The Nuances of GOFO
Gold message boards have been aflutter in recent weeks with GOFO rates once again dipping into negative territory. The gold bugs' fire was further fueled when an article clumsily titled "Strange GOFO cry heralds trouble for gold" was published in the credible Financial Times, as opposed to the internet backwaters where articles on the subject usually appear. Clumsily titled because it goes on to explain why negative GOFO is actually good for physical gold, but trouble for paper gold. We're here to say that GOFO rates dipping back into negative territory doesn't herald much of anything, but rather reinforces underlying themes that have been in place for some time.
Before getting into the nuances of GOFO, it's important to understand what it represents. GOFO is simply the forward rate for gold - hence the name GO(ld)-FO(rward). You can think of it almost like an interest rate for an asset with no yield. All commodities have forward curves that represent the price at which you can buy a fixed amount in the future. The forward prices for various durations create, in effect, a yield curve which is shaped by many factors including supply/demand, storage costs, etc. Unlike consumable commodities, like corn or crude oil for example, gold functions as a unit of exchange, except that unlike cash it doesn't earn interest. However, since gold is as good as (and sometimes preferable to) cash, GOFO rates typically mimic LIBOR rates. Gold can be expensive to store, estimated to be roughly 10bps for 3 months, so in theory GOFO rates should always be slightly below LIBOR. As soon you can see on the chart below, that's not always the case but we'll come back to this later.
3m GOFO (orange) - 3M LIBOR (white)
If nominal interest rates in the US have never been negative, and GOFO is supposed to trade near LIBOR: why have GOFO rates been negative twice in the past four months? Negative GOFO rates imply that buyers are willing to pay a premium for immediate delivery of gold. Historically, gold miners looking to hedge forward production would flatten the forward curve, but miners haven't been hedging for quite some time. Beyond that, with prices down on the year, it's hard to imagine that the demand for gold is so strong that buyers would be willing to pay a premium for immediate delivery. In our view, the huge selling of paper gold / gold derivatives and the unprecedented demand for physical gold from China have resulted in the anomaly of negative GOFO rates. These are not new themes, but do give important clues about what is really happening in the opaque market for gold.
On July 9, 2013 the CFTC's Commitment of Traders report showed a gross short position of 17.7m ounces - far and away the all-time high dating back to 1995. 17.7mn ounces, or 551 tons, of gold at $1,300/oz is roughly $23bn USD. For comparisons sake, that is more gold than the official reserves of the UK, Sweden, or Saudi Arabia. A noteworthy short position sure to draw the ire of gold bugs everywhere? Yes. A sign of manipulation? No. In fact, on July 9th, the CFTC net speculative position was still long 3.9mn ounces, although that was the smallest net long position since January 2002.
This massive short position was accumulated trading the August 2013 gold future contract. However, by the end of July that contract came due for delivery and traders had to either produce a lot of physical gold or roll over their short position to the December 2013 contract (where the liquidity was shifting). Considering that the price of gold fell roughly $370/oz. between May 11th - July 9th and the vast majority of speculators have no ability to deliver physical bullion - rolling over their existing short position became the obvious choice.
When a short futures position is rolled forward there are actually two transactions that must take place. The existing short must be covered, or bought back, and the forward contract is sold, putting pressure on the forward curve to flatten. When a historically large short position is rolled in a relatively short period of time, there is a lot of downward pressure on the forward curve and some funky things can happen. This is what drove GOFO rates negative, and the chart below shows that GOFO rates remained depressed well after the peak CFTC short position (as the contract rolled).
CFTC Non-Commercial Shorts (orange, inverted) - 3m GOFO (white)
This activity is not nefarious. The largest short future position in CFTC history amounted to $23bn, but the World Gold Council estimates the investable gold market is nearly $2.5trn. The often cited ETF holdings currently amount to 64mn ounces, $83bn, or just 3.3% of investable gold. That such a small volume of gold can cause such a disruption in prices is not manipulative or even a new phenomenon, just surprising. Subsequently, the level of CFTC shorts declined and GOFO rates started functioning normally again… until recently.
On October 10th, the 3-month GOFO rate was trading around 13bps but suddenly dropped to -4bps over the next 6 days. At the time, there were rumors of shortages in 400oz. good delivery London bullion bars, the very bars used to back all major gold ETF's. Shortages in specific products are not a huge problem because gold is malleable and can be melted down into any form, but the process is not immediate. In the meantime, investors appeared willing to pay a premium for spot gold - once again creating backwardation or a downward sloping curve. What happened to all the 400oz. bars? It looks like they've been going to China in astounding quantities.
Reports have circulated that this year's UK gold exports to Switzerland have increased nearly 10 times Y/Y. Through the first six months of the year, the UK has sent 798 tons ($33.5bn) of gold to Swiss refineries compared to 83 tons in 2012. In the month of May alone, 240 tons were exported. Most of these exports were sent to Swiss refineries, where an estimated 70% of the world's gold supply is processed. There are also unconfirmed reports that these refineries are working around the clock at full capacity to handle the enormous flow.
Swiss refineries are the only ones in the world that have the capacity to transform London Good Delivery 400oz. bars into smaller bullion products, like the 1-kilo bar, that Asian buyers prefer. Import data tracked by the Hong Kong Census & Statistics Department shows that 375 tons of gold have been imported from Switzerland over the first six months of this year. In addition, the UK sent another 59 tons directly to Hong Kong, and 22 other countries exported a total of 135 tons over the same time period.
These totals are staggering. If a made-up country started from scratch and imported 807 tons of gold, they would have the 8th largest official holdings in the world. Again, this is not a new theme. China has been expected to overtake India as the world's largest gold consumer for a while now, but the scale at which gold is being sent to China is causing strange reverberations in the gold market, resulting in confounding statistics like negative GOFO rates. Since bottoming at -4bps on November 1, rates are now back into positive territory.
Gold entering into backwardation (where immediate demand > supply) should be hugely bullish for gold, however, the anomaly has been largely an explainable one-off. The gold price has gone up with GOFO below zero, but not much.
3m GOFO (orange) - Gold Price (white)
Although negative GOFO rates don't get us particularly excited, there is an even more obscure metric that is far more important to us: the gold lease rate. The gold lease rate is simply LIBOR minus GOFO. In other words, it's the rate at which gold can be borrowed. Back in the 1990's the gold lease rate market was much more liquid and active when gold miners hedged their forward production. However, since peaking in 1999 the global mining hedge book fell dramatically as miners looked to increase their exposure to rising gold prices.
3m Gold Lease Rate (orange) - Global Net Miner Hedging (white)
Gold miners slashing their hedge books was a serious hit to demand for gold borrowing, and consequently, lease rates fell close to zero. Except for bullion banks, which specialize in borrowing/lending gold, the fundamental shift in lease rates had little impact on the markets, but did significantly alter what the rate represents. Today, the lease rate is more of a systemic risk barometer. In fact, since the early 2000's, lease rates have had a strong correlation with a true barometer of systemic risk: US swap spreads (the difference in return from lending money to a bank vs. the US government: "the risk-free counterparty" - chart below).
As mentioned earlier, in a vacuum, LIBOR should always be roughly 10bps (for storage) above GOFO for a duration of 3-months. Unfortunately markets don't operate in a vacuum, and in 2008 during the Lehman bankruptcy lending markets froze. Banks wouldn't lend to each other, and gold (the ultimate form of collateral) became a much more valuable asset to have on a balance sheet - increasing the cost to borrow gold dramatically.
3-month Gold Lease Rate (orange) - US 2Y Swap Spread (white)
Lease rates stayed elevated through the first quarter of 2009, coincidentally the stock market bottomed at the same time, when the Fed inundated markets with liquidity and counterparty risk dropped. With counterparty risk no longer an issue, banks quickly began raising cash reserves in order to shore up their balance sheets. This pressure for bank liquidity was especially acute in Europe during the sovereign debt crisis. One of the easiest ways for banks to raise capital was lending out their gold. With all the banks trying to do this at the same time, it created a surge of gold available for borrow with little increase in demand. This ultimately pushed lease rates into negative territory, creating the perverse situation where traders were paying for another counterparty to borrow their gold. Now that balance sheets have been sufficiently capitalized, the gold leasing market has normalized with 3-month rates now trading around 15bps.
What does all this mean for the directional price of gold? Not much. When the lease rate shot up 300 bps during the Lehman crisis (theoretically very bullish), gold actually traded lower. And conversely when lease rates were negative (theoretically very bearish), the gold price more than doubled from 2008-2011. The lease rate, although no longer negative, is still trading at very depressed levels signifying little systemic risk in the financial system. Banks have stronger capital reserves than 2008 and interest rates are still at the zero-bound, leaving little interbank counterparty risk. But it's telling that when lending markets did freeze up, the attention immediately shifted to gold. Gold is one of the few assets without a counterparty making it invaluable during a crisis.
GOFO and lease rates acting irrationally can make for alarming headlines, but at it's always best to keep things simple. Physical gold has no counterparty, has been battle-tested for hundreds of years, and is the ultimate form of collateral. The paper gold market will thrash around causing distortions, and unallocated gold will be imprudently leased across the globe, but allocated physical gold always maintains its intangible value. In the midst of increasingly complex and globalized financial markets, turmoil is almost inevitable. The key is to prepare before the storm, not during it, and an allocation to gold is a necessity as financial markets roll through unchartered territory. If/when the next crisis does appear banks won't be looking to borrow ETF's or futures; they'll want the real thing: allocated physical gold. And you should too.