Why gold forward rate inversion is important | FT Alphaville

Why gold forward rate inversion is important

Here’s a crazy situation to consider.

The gold lease rate (which can also be understood as gold Libor, the gold interest rate or the cost of shorting gold) is becoming increasingly negative at the short end. This is the natural consequence of Gofo rates rising ever more greatly beyond Libor costs at the front end. Since the gold lease is derived from the calculation of Libor minus Gofo, any instance where Gofo is greater than Libor leads to a negative gold lease rate.

This, for example, is the historic path of the three-month gold lease rate:

And here’s a closer look at the near-term action:

See, there’s been a collapse.

Yet the irony is that there is still a cost to borrow the GLD exchange traded fund. Why should it cost you to borrow GLD at all, but earn you a return to borrow gold (Ed- surely administrative costs)? Weren’t the two supposed to be a like-for-like? What’s more there have been very interesting bursts of short-interest in GLD recently, according to Data Explorers.

But first back to gold lease rates.

Gold lease rates first went negative (in this recent spell) in March 2009, and have remained almost consistently negative since about July of that year (with brief interim spells of positivity). But it’s only in the last week that the rate has become severely and almost illogically depressed.

Interestingly — despite talk of funding pressure all round — this doesn’t tie with the scenario experienced during Lehman at all. At that time gold lease rates spiked, reaching historic highs. In fact, it was only with the onset of extraordinary liquidity in November that lease rates began to fall quickly.

Before that happened, something extraordinary transpired. Gold forward rates flipped, ever so briefly, into backwardation. A situation which has hitherto only really been seen in the Japanese gold market (where gold is denominated in yen).

Of course if you consider gold the soundest money carrying the lowest interest rate structure — the ultimate risk-free rate — it makes sense that gold lease rates should closely echo market interest rates, but always remain fractionally lower. So when Libor shot up during the crisis, gold lease rates understandably followed behind them — leading to a situation where if you had gold, you could lend it out for a very high return, since the risk lay with holding cash on reserve rather than gold and you had to be compensated. Gold was the ultimate collateral. Or in other words, the demand for gold rose alongside the expense of borrowing in the interbank market.

Eventually the demand for gold became so great that the lease rate overshot Libor, leading to the backwardation discussed above. At that point gold became a bit of a Giffen good. It was — if you believe the goldbugs — a situation which possibly signified the death of paper-money.

But even then, while gold itself became backwardated, the Gofo curve remained normal, as did the Libor curve.

What we have now, however, is quite the opposite. Gofo remains in contango — i.e. continues to avoid backwardation — while gold lease rates have fallen sharply into negative territory instead. Almost as a counterbalance.

On the surface, the negative rate implies extremely low demand for gold compared to cash. Or you could say, there are currently more people prepared to lend gold at a terrible rate (because there’s so much of it around) than prepared to lend their cash for gold. (Contrary to anecdotal reports from the physical market which suggest a lack of gold sovereigns and such the like.)

In this scenario, nevertheless, gold is seen as the risk. Gold has become a lousy monetary substitute, and is anything but optimum collateral. In fact, it is US Treasuries that are being over-bid, not gold.

The lower the rate goes, the more it suggests an extreme rush to pawn gold in exchange for cash in the marketplace. Anything but gold, you might say. Meanwhile, the more gold that ends up at the pawn shop, the less favourable the rate received for pawning in the market. (The pawn shop doesn’t want to carry all that risk and has to cover that risk by offering less cash for gold.)

Thus it’s not money that is dying, quite the opposite, it’s gold.

But there is one important other factor to consider: Central bank intervention.

Why on earth would anyone be prepared to lend gold at a negative rate for almost two years, when demand in the physical market was supposedly so huge?

If you consider that these actions were what prevented gold from flipping into backwardation, perhaps it becomes a little clearer.

As Reginald Howe at the Golden Sextant points out:

At the LBMA, therefore, gold continues to avoid backwardation, but only because central banks continue to lend at historically very low lease rates. It is a strange situation. Gold for spot delivery and bullion funds with high credibility for physical possession of metal in the amounts claimed are selling at premiums over paper of lesser reliability. But where gold is arbitraged against currencies on the basis of relative interest rates, it remains in contango.

Now, if this is the result of some sort of central bank intervention, it’s clear that the central banks still have to find end demand for that lent gold (via the bullion bank intermediaries). The gold mining companies that used to be  counterparties, have closed their hedging books. They’re no longer willing takers of borrowed gold.

Who’s the only viable candidate left? Answer: Gold ETFs and gold exchanges.

While the likes of GLD insist every share outstanding is matched by a gold bar — and this is almost definitely true — what they can’t claim is that there’s a way to differentiate gold with previous claims on it from gold without previous claims on it within its reserves (i.e. borrowed gold). It is consequently entirely possible that gold ETFs are sitting on mountains of borrowed central bank gold.

GLD’s defence, of course, is that prior claims on its reserves are not their concern. They are the liablity of the party that delivered the gold to GLD (almost certainly a bullion bank). Thus it is the bullion bank that risks being squeezed on delivery in the physical market, not GLD.

Of course, with a negative interest rate for borrowing gold, the bullion banks are being more than compensated for the risk of a squeeze. On top of everything they can always create new GLD shares ad infinitum, if needs be. (At least until all central bank gold stock has been lent into gold ETFs, arguably forcing central banks to replenish the gold lending pool via market purchases.)

In the above scenario — in which bullion banks are possibly recycling borrowed gold into GLD shares to sell into the market — these short sales will put pressure on GLD units themselves.

As Howe noted earlier this year:

In recent months, while GLD has generally sold at a slight discount to net asset value, other bullion funds with more transparent and credible custodial and auditing procedures have commanded significant premiums. E.g., Central Fund of Canada (CEF), Central Gold Trust (GTU), Sprott Physical Gold Trust (PHYS). Anecdotal evidence also continues to surface of premiums for spot delivery of physical metal or cash settlement in lieu of physical.

Thus, if central banks are really using bullion banks as feeders of borrowed gold into gold ETFs (so as to suppress prices and keep gold lease rates negative, avoiding gold backwardation), you’d assume the strategy could easily unravel if and when people started liquidating large portions of GLD, i.e. forcing delivery of that gold.

Gold and the create-to-lend mechanism

Ordinarily when new shares are created in an ETF for shorting purposes, there is no respective spike in assets under management. That’s because shares are often created using the “create-to-lend” facility, which sees shares issued against borrowed stocks which are almost immediately sold into the market. Since there is no incremental demand for those shares, an oversupply of units hits the market place causing the ETF arbitrage mechanism to kick in. The very same shares are thus almost immediately redeemed, leaving assets under management unchanged but the short-interest ratio higher.

This, by the way, is how we get to such large short-interest ratios in some popularly shorted ETFs.

Of course in the bullion scenario, one could argue that the shorts are continuously lapped up by strong demand from GLD buyers. The shorts are thus disguised by continuing assets under management growth — a fact which leaves the short-interest ratio relatively stable, and has a slowing impact on AuM growth if anything.

But if a large GLD share owner coincidentally liquidates a sizeable portion of GLD shares one day (in order to take delivery of real gold instead), this could theoretically destabilise the short-interest balance. A fact which may or may not have happened recently.

As the short-interest data firm Data explorers noted at the end of August:

The amount of money invested in SSGA’s Gold Trust (GLD) is very close to the total assets in the instrument, which tracks the S&P 500 (SPY) at USD 71bn. GLD issues shares in exchange for deposits of gold. On August 10th, there was an unusually large rise in short selling by 250% to 18m shares. This position was then immediately covered as gold began its recent ascent to $1,917.9 on Monday of this week.

There has been little activity in this ETF’s other listings in Hong Kong, Singapore and Tokyo. Another large and physically backed (i.e. owns gold) ETF is the iShares Comex Gold Trust (IAU). This has sporadic spikes in short interest, and these spikes have increased in their frequency over this past quarter, showing some evidence that the need to hedge or short gold is more frequent than it was.

If we look at the number of securities lending transactions, we see a 20% increase in the number of loans in GLD in the last week alone. The absolute rise in shares short might not be that cataclysmic, but more trades might well mean more positioning, which translates to nervousness that the gold price cannot sustain such lofty levels. The equivalent rise in trades in IAU is 84% since last Thursday – a huge change.

The fact that the short position was immediately covered suggests that whoever liquidated GLD sold the bullion back into the market. By doing so it allowed the short parties to cover their position, bringing the overall short-interest ratio back to equilibrium quickly.

Whatever the case, one thing’s for sure. Volatility in the gold/GLD spread (blue line below) has increased since August 10, which was about the time of the Paulson GLD liquidation rumours:

So has the GLD liquidation possibly disturbed the short-ratio balance? Is this why gold lease rates have had to adjust radically downwards? Are bullion banks demanding increased compensated for supplying borrowed gold into gold ETFs because there’s suddenly been a genuinely large amount of gold thrown back into the gold system?

Is this also why AuM has stagnated and GLD has become ineffective as a central bank policy tool?

Who can say, but all these factors are definitely worth thinking about we would argue.

Related links:
What do silver lease rates have to do with fund redemptions?
– FT Alphaville
Gold Derivatives: GLD and Ass Backwardation
– FT Alphaville
Cash for gold, financial market edition
– FT Alphaville
The secured lending boom (through gold-tinted glasses)
– FT Alphavill