Gold deposit rates - a guidance paper

Background

Gold’s intrinsic value has been recognised for millennia but its role on the global stage changed fundamentally following the collapse of the Bretton Woods system. 

In 1971, US President Nixon suspended the dollar’s convertibility into gold. By 1973, advanced countries began adopting a managed floating exchange rate system and gold started to trade freely on world markets.

Confronted with a new international monetary system that no longer had ties to gold, economies around the world re-examined the need for gold as a reserve asset. Many major holders of gold reserves, primarily Western European central banks, decided to sell or reduce their holdings. Between 1971 and 2009, central banks made net sales of 8,388 tonnes of gold.⁹ At the same time however, a more active gold lending market evolved, as central banks strived to leverage their vast gold holdings and generate returns beyond capital gains. 

There are two ways in which a central bank can generate yield from its gold holdings. The first is through the uncollateralised lending of gold, most commonly to a bullion bank. Given that gold is a monetary asset for central banks, it can be lent out on term deposit in the same way as any other currency in their reserve portfolio. Typically, a central bank will place gold on deposit with a bullion bank, in return for a deposit rate. Maturities vary but the most frequent are 1-month, 3-month and 12-month tenors. At the end of the term, the gold is returned with the interest paid either in gold or currency, although the bars will be different from those originally lent out.¹⁰ Deposit rates are derived and set independently by bullion banks.

Central banks are also able to generate yield from their gold holdings via a gold swap, or more specifically, a repurchase agreement that simulates a swap. In this transaction, a central bank sells its gold to a bullion bank with the promise to buy back the gold at a later date. The central bank pays interest equivalent to the GOFO rate so, in this context, the GOFO rate could be understood as a US dollar loan using gold as collateral. Officially, it is defined as the rate at which market-making members¹¹ of the London Bullion Market Association (LBMA) will lend gold on swap against US dollars.¹² The central bank is then able to reinvest the funds at Libor and earn the spread between Libor and GOFO, which in this case will amount to the gold lease rate. 

As an independently derived, over-the-counter instrument, gold lease rates can be best understood through the interaction of the demand and supply of borrowed gold. When the market is oversupplied relative to demand, gold lease rates will fall. Conversely, when the market is undersupplied relative to demand, gold lease rates will rise.

The gold carry trade

It is hard to examine gold lease rates without first understanding the gold carry trade. The gold carry trade once played a significant role in the gold lending market and constituted a large part of the demand to borrow gold. However, as market conditions changed over the years, the risk of the trade increased while its profitability declined. This reduced demand for borrowed gold and may be one of the leading causes behind the decline in gold lease rates. 

Under the gold carry trade, a trader, typically a bullion bank, hedge fund or other financial institution, borrows gold at the gold lease rate, sells the gold and invests the funds in higher-yielding instruments. At loan maturity, the trader sells the higher-yielding asset and buys back the gold to close its position with the lender. The ‘carry return’ of the trade is the spread between the higher-yielding asset and the gold lease rate (the rate at which the gold  is borrowed). 

If the trader invested the funds at the London Interbank Offered Rate (Libor),¹³ the ‘carry’ will be the GOFO rate. Put another way, the gold lease rate can be derived from the difference between Libor and GOFO.¹⁴

Derived gold lease rate = Libor – GOFO

The gold carry trade incurs one major risk. As the trader has effectively sold the gold short, an increase in the price of gold at loan maturity would result in a loss of profitability. The gold carry trade thus only makes sense when the price of gold is expected to remain stable or decline. 

If the difference between US dollar interest rates and the cost to borrow gold (represented by Libor and the derived gold lease rate above) is wide, the carry return appears more attractive and there may be an increased uptake 
of the gold carry trade, leading to more gold borrowing demand and higher gold deposit rates for central banks. 

The gold carry trade proved to be a particularly profitable strategy during the late 1990s, when gold prices were falling and higher-yielding assets, in particular yen-backed securities,¹⁵ were rising. During this time, the gold carry trade added supplies to the market and gold prices fell, reinforcing the negative sentiment towards gold. Gold lease rates were volatile, but also elevated. 

Market conditions changed however after the CBGA in 1999. This was followed by the 9/11 terrorist attack in 2001, the bursting of the dot-com bubble, and, from 2001, the end of a protracted period of US dollar strength, all of which contributed to more positive price expectations for gold. Against this backdrop, carry trades incurred greater risk and frequently proved unprofitable.¹⁶ As such, both the demand to borrow gold and gold lease rates fell. 

It is worth noting that speculators also use the gold borrowing market to bet on price declines through the sale and subsequent repurchase of gold at a lower price. When supplies dry up and speculative pressure rises, gold lease rates can jump. This has happened in the past and could, of course, happen again, particularly if sentiment turned negative on gold.¹⁷

Gold producers’ hedging demand

Gold producers’ hedging constitutes a large proportion of the demand to borrow gold and is thus another critical determinant of gold lease rates. During gold bear markets, there are two common approaches producers can use to hedge their exposure to the price of gold. 

First, before the gold is refined and ready for sale, the producer can borrow gold from a bullion bank, sell it at a higher price today and use the funds to finance ongoing expenses. At the maturity of the loan, the producer can close its position with the bullion bank with the production of its gold. Gold loans were historically a common instrument used by mining companies but have become less prevalent as banks adopted more stringent credit criteria. 

Second, the producer can enter into a forward sale agreement with a bullion bank and lock in the selling price of its gold. As the risk of a decline in the price of gold is passed onto the bullion bank in the forward sale agreement, the bullion bank will have to offset its exposure by borrowing gold and selling it in the spot market. The funds raised from the sale are then invested in the short-term interest rate market. Upon the expiry of the forward contract, the bullion bank closes its position via the delivery of gold it receives from the producer. 

Figure 1: Mechanics of a forward sale agreement between a bullion bank and a producer

Both approaches involve the borrowing of gold. In fact, it is common for banks to insist that producers hedge their exposure to the price of gold before agreeing to lend them capital for infrastructure development and acquisitions.¹⁸ 

There are three distinct periods in the history of gold producer hedging activity. Each of these had a specific and marked effect on gold lease rates.

Discovery and ramp up: The first period ran from the mid-1980s until 1999. During this time, the US dollar gold price was broadly flat or in decline, while interest rates in US dollars and other producer currencies, in particular the Australian dollar, were high. Low spot gold prices and high interest rates created large forward premiums. And gold miners took advantage of these conditions by selling production forward using simple forward sales and more complicated options-based structures. As a result, the aggregate producer hedge book grew steadily during this period, reaching a high of 3,231t in 1999.¹⁹ Gold lease rates rose concurrently, with the 12-month rate averaging 1.39% over the same period. 

As the price of gold fell from around $370/oz in the mid-1980s to a low of $252/oz in 1999, producer hedging increased. But conditions changed after the first CBGA in 1999, and were compounded by the 9/11 terrorist attack in 2001. The gold price rose sharply, several gold mining companies were hit by very large negative mark-to-market positions on their hedge books and some companies had to be financially restructured as a consequence. 

This experience had long term consequences for producer hedging: public anger from shareholders about the decline in some gold mining company share prices due to a rising gold price led to an end to extensive gold producer hedging practices. 

Global hedge book buy-back: From 2000 to 2010, the aggregate producer hedge book declined each year as companies delivered into their positions and, in some cases, closed out or ‘bought back’ their gold hedges. By 2010, the book had fallen to 300t and it has not materially increased since. During this time, gold lease rates broadly declined.

A return to limited hedging: Some producers returned to the hedging market in recent years, when the price of gold was in decline, following the record high of $1,921/oz in September 2011.20 But positions were both smaller and shorter in duration than in the 1980-1990s. 

Interest rates and gold lease rates

The opportunity cost of holding gold, represented by the prevailing level of interest rates, has often been considered as an important driver of the price of gold. The onset of the global COVID-19 pandemic in 2020 shone a light on this relationship when the fall in real interest rates, among other factors, led to a strong outperformance in the price of gold. 

The link between real interest rates and the price of gold historically also has extended to gold lease rates.²¹ When real interest rates trend lower, gold price sentiment improves, and gold borrowing demand stemming from producer hedging, the gold carry trade and speculative short-selling falls, leading to a decline in gold lease rates. Existing research in this area has, in fact, argued that gold lease rates can been seen as a proxy for real interest rates²² and our model supports this conclusion to an extent, with the US 10y TIPS flagged as an important variable in explaining the behaviour of gold lease rates.

In general, market interest rates, such as Libor, should also serve as a theoretical ceiling to gold lease rates. Unlike fiat currencies, gold cannot be debased through monetary policies and has no political or counterparty risks.  

As it commands less risk than fiat currencies, it should,  in theory, offer interest rates lower than those for US dollars or other currencies. In addition, the long-term growth in gold supply historically has always been slower than that of any national currency, which further explains why gold usually bears a lower interest rate structure. 

Meanwhile, there are also sufficient above-ground  stocks of gold, mainly from central bank gold reserves, to prevent gold lease rates from rising above US dollar interest rates.²³

In reality, however, there have been occasions when the gold lease rate was higher than Libor, gold forward rates turned negative and the market moved into backwardation.²⁴ At such times, owners of gold are incentivised to sell gold at spot and buy forward at a lower price to pocket a profit on the arbitrage. This will push down the spot price of gold and allow the gold forward curve to eventually return to contango. 

Dollar funding demand

The gold lease rate can also be affected by broader market developments and events, most recently the coronavirus pandemic. During the first quarter of this year, as the pandemic intensified, policymakers around the world imposed restrictions on the movement of goods and people and millions of businesses closed their doors. 

Stock markets initially tumbled as investors liquidated financial assets to raise US dollars and move into the safety of US treasuries. Corporations also scrambled to ensure they had sufficient US dollar liquidity to tide through the demand shock caused by the pandemic. 

Meanwhile, the crash in oil prices also slowed the flow of US dollars to oil and commodity producing countries and exacerbated the stress on US dollar funding. During this time, the premium paid to swap Euro for the US dollars widened to the highest level since December 2017.²⁵ 

This, in turn, affected the gold lending market. The 3-month gold lease rate rose into the positive and diverged from both EURUSD and USDJPY cross-currency basis swaps as banks and corporations sought to raise funding by borrowing gold and swapping it into US dollars. Elevated rates persisted until the US Federal Reserve initiated US dollar swap lines with central banks around the world. US dollar liquidity abruptly eased and the 3-month gold lease rate briefly fell into negative territory (although it rebounded within 24 hours along with the cross-currency basis swaps).

A similar situation occurred before and during the Global Financial Crisis. Markets were jittery well before Lehman Brothers collapsed in September 2008 and both Libor and GOFO fell in the early days of the financial crisis. However, deteriorating creditworthiness and cash hoarding by banks to meet liquidity obligations²⁶ caused the decline in Libor to be comparatively subdued. As a result, the spread between Libor and GOFO widened and gold lease rates started to rise. 

Inventory building demand

Gold refineries are also large users of borrowed gold. Typically, they must pay for feedstock shortly after it is delivered to the refinery, but only get paid themselves when the gold has been refined, fabricated into bars or other forms, and sold. When demand for refined products increases rapidly, refineries need to increase gold borrowing to fund their larger work in progress. 

In June 2013, for example, the price of gold fell below US$1300/oz, prompting a flurry of demand from Asia. Gold had risen to US$1900/oz just a couple of years previously and Asian buyers found the lower price a bargain. Greater China’s demand for gold bars and coins rose to 158.8 tonnes in Q2 2013, the highest quarterly demand on record.²⁹ Between May and June 2013, Asian retail investors paid an average premium of US$30 an ounce and as much as US$56 an ounce to secure physical gold products.³⁰ 

Seeking to capitalise on this demand, dealers started to draw down inventories in US COMEX³¹ warehouses to recast into LBMA specifications or kilobars of 9999 purity favoured by Asian retail investors.32 The 1-month gold lease rate subsequently rose to around 15 basis points, the highest since the 2008 financial crisis, as additional borrowing was required to fund the additional refining  and fabrication.

Central banks and the supply of gold available for lending

As the primary lenders of gold, central banks play a key role in the gold lending market and their lending activities have been found to lead to lower gold lease rates, higher forward prices and increased hedging demand from producers.³³ 

This was graphically illustrated with the announcement of the first Central Bank Gold Agreement (CBGA ) in 1999. Also known as the Washington Agreement on Gold, this was arguably the most pivotal supply-side event in the gold lending market. 

During the 1980s and 1990s, central banks were noted sellers of gold, a trend that became increasingly marked in the run-up to the millennium. As this coincided with growing demand for borrowed gold, many Western European central banks also extended their use of lending, swaps and other derivative instruments. An increase in lending typically resulted in additional gold being sold, meaning that the trend was adding further supplies to the market.

Persistent selling activity destabilised the market, while broader fears about central bank intentions prompted further falls in the price of gold. This created a prolonged bear market, which in turn encouraged gold producers to take out hedging agreements with bullion banks, ultimately creating an environment where gold lease rates were elevated and volatile. 

This period came to an abrupt end with the first CBGA on 26th September 1999. Under the Agreement, 15 European central banks agreed to limit their collective sales to 2,000 tonnes annually and to maintain or reduce gold lending and derivative activity over the following five years. News of the Agreement prompted a sharp spike in the price of gold. The gold lease rate also jumped, amid expectations that the amount of gold available for borrowing would decrease. The 3-month gold lease rate rose to approx. 9% on the 29th of September 1999, the highest on record. However, it fell sharply in subsequent months, slipping below 1% within a year of the CBGA (Chart 10). 

Conditions have changed significantly since then. Central banks have become net buyers of gold, the sources of demand have become more diverse and, in recent years, sentiment towards gold has become increasingly bullish. Against this backdrop, gold lease rates have remained subdued. Nonetheless demand for return persists. In the World Gold Council’s 2020 Central Bank Gold Reserves Survey, 47% of those surveyed said that they are actively lending out their holdings, while 53% said that they are involved in swap transactions with their gold reserves.³⁴

Other theoretical drivers of gold lease rates

The convenience yield 

The convenience yield of gold is the implied return from holding gold on inventory. It usually increases when there is a strong possibility – perceived or actual – that future supplies will be disrupted. As such, it can be seen as the benefit accrued from hedging against supply disruptions. 

The convenience yield also plays an implicit role in the forward market, insofar as the forward price of gold could be defined as its spot price plus the cost of financing and storage minus its ‘convenience yield’. 

Under normal circumstances, the forward price of gold is higher than the spot price, in part reflecting the carry cost of gold. In other words, the gold forward curve is in contango. At times, however, the forward price of gold is lower than its spot price, so the gold forward curve is in backwardation. This can occur when the physical demand for gold is high and the convenience yield of holding gold outweighs the twin costs of financing and storage. In general, a higher convenience yield leads to lower forward rates and thus a higher gold deposit rate. This is supported by existing research on gold lease rates, which finds the convenience yield is well approximated by the gold lease rates.³⁶ 

The COVID-19 pandemic illustrates this point. As planes were grounded and Swiss refineries closed, traders struggled to deliver against future contracts traded on COMEX that required 100-oz bars or kilobars. Their difficulties arose just as gold’s safe-haven status provoked worldwide investor demand. Under normal circumstances, traders would have been able to satisfy this demand by procuring 400-oz bars traded in London, recasting them into 100-oz bars or kilobars in European refineries and delivering them against future contracts in New York. In the early days of the pandemic, however, this proved either impossible or too costly. As a result, the divergence between New York futures and the London spot price increased to unprecedented levels.³⁷

During this time, the convenience yield was also elevated, as sellers of gold scrambled to secure sufficient supplies. The gold forward curve temporarily moved into backwardation before reverting to contango in recent months. At the same time, gold lease rates rose sharply, with the 3-month rate moving from negative territory to almost 1%.

Conclusion

Gold lease rates are governed by diverse factors, some pertaining directly to the gold market and some to wider macro-economic and geopolitical trends. Rates are also independently set on an over-the-counter basis by bullion banks and may vary from bank to bank. 

As such, they can be volatile and unpredictable. Nonetheless, in-depth analysis of past events, coupled with our proprietary model, offers useful guidance about the direction of gold lease rates and the drivers that influence them. Further, our model has been able to provide statistical validity to back up the theoretical and empirical impact of the following factors on gold lease rates. In decreasing order of importance,³⁸ these factors are:

Real interest rates: Interest rates represent the opportunity cost of holding gold. When real interest  rates are trending lower, the gold price sentiment will improve and the demand to borrow gold for the purposes of hedging, speculative short-selling and the gold carry trade will decline. This will lead to a fall in gold lease  rates. In addition, low real interest rates will also result in smaller carry returns and render the gold carry trade less appealing. 

Central bank sales: When central banks are active sellers, this tends to impact gold market sentiment and support higher lease rates. Central bank gold selling is often linked to gold lending activity too, providing the means by which investors can conduct gold carry trades and producers can take out hedging instruments. Since 2010, central banks have, however, been net buyers of gold. This has improved sentiment in the gold market and led to lower gold lease rates. Specifically, gold holdings of central banks that are signatories to the CBGA have proven to be a good predictor of gold lease rates, probably because western central banks hold more gold and have historically been more active in the gold lending market.

Producer hedging demand: This is a key determinant affecting gold lease rates, in turn affected by gold market sentiment. When producers fear that gold prices will fall in the future, they are more inclined to take out forward hedging agreements so they can better manage future production. That leads to greater gold borrowing demand and tends to drive gold lease rates higher. In recent years, due to better gold market sentiment and shareholder backlash from aggressive hedging activities seen in earlier years, producer hedging demand has fallen, bringing gold lease rates along with it.

One-off crises: During periods of financial or geopolitical turmoil, such as the Global Financial Crisis or the coronavirus pandemic, demand for gold tends to rise significantly, as investors seek out safe-haven assets. At the same time gold borrowing demand also increases, as financial institutions and corporations look to borrow gold to swap into US dollars in order to bolster their liquidity. At such times, equity prices tend either to fall or become increasingly volatile and gold lease rates invariably rise. These one-off crises usually see an increase in the real price of gold as well as an increase in the annualised volatility of the S&P 500 and both variables have been flagged as important in our model. 

Price expectation of gold: Producer hedging demand and the attractiveness of the gold carry trade are ultimately linked to the price expectation of gold. If the price expectation of gold is bullish, producer hedging demand will fall and the gold carry trade will become less lucrative. This will lead to lower gold lease rates. However, the price expectation of gold is in turn dependent on, amongst other factors, central bank selling intentions, real interest rates and market volatility. As a result, speculative positioning  in gold, as represented by the CFTC net positioning, has a role in explaining how gold lease rates will evolve  in the future. 

Today’s environment is characterised by an increasingly bullish sentiment towards gold, lacklustre hedging demand and a steady increase in central bank gold buying. Against this backdrop, gold lease rates have been low. However, as the recent pandemic indicates, when conditions change, gold lease rates follow suit – and the shift can  be both marked and rapid. 

More details on data, model approach and results can be found in the appendix in the full report.