The Gold Slam Is Evidence Of Troubled Banks
Back in July 2010, the Wall Street Journal caused some commotion when it happened to notice in the annual report for the Bank for International Settlements the sudden appearance of gold swap operations to the tune of 346 tons. Subsequent investigation by media outlets, including the Financial Times, reported that the BIS had indeed swapped in 346 tons of gold holdings from ten European commercial banks. That was highly unusual in that gold swaps are typically conducted between and among central banks.
Included in that list of commercial banks were, according to the Financial Times, HSBC, BNP Paribas and Société Générale. The timing of the swaps was pinned down to sometime between December 2009 and January 2010 - just as the world was getting reacquainted with the Greek Republic.
Through the "normal" course of its operations, the BIS had accumulated a large stockpile of US dollar currency "reserves" in its accounts up to that point. As the sovereign debt crisis began to flare its initial heat, bestowing a growing mistrust onto certain commercial banks, liquidity in the wider banking once again agonizingly withdrew. The Great Financial Crisis was only a few months in the past, and yet it seemed as if the great central bank measures of the prior year would not be enough.
Under these circumstances, the BIS initially reported that the gold swaps were its "normal course of business" and that there was nothing at all unusual in undertaking them. However, in its annual report of March 2009, the BIS had reported $0.00 of "gold derivatives". By that March 2010 annual report, the one released in June 2010 that started the confusion, "gold derivatives" totaled $8.162 billion. Gold "deposited in BIS sight accounts" rose from $23 billion to $32 billion. There had been no other indications of swap agreements going back to at least 2000.
To say this was the normal course of business did not hold much logic or sense in light of any reasonable analysis. In fact the opposite was rather clear, particularly in the circumstances as they related to European banks in early 2010 - these European banks were in dollar trouble and needed any outlet they could find. Their holdings of PIIGS sovereign debt were being rejected as usable collateral through regular repo funding channels, and the banks were increasingly running out of their margin of safety in funding terms. The ECB had not yet begun to even think about its emergency workarounds until this gold swap had actually been publicized.
The question of the swaps morphed from interest in commercial banks' short dollar positions to where all this gold actually came from. It was conventional wisdom at the time that the only banks that actually had access to such quantities of gold were central banks. So, logically, fingers began to point toward sovereign gold stashes in the rush to find which country was "bailing out" its banks through gold.
But in an article dated July 29, 2010, the Financial Times offered an alternate, and intriguing, explanation:
"Investors have bought physical gold in record amounts during the past two years and deposited it in commercial banks. European financial institutions are awash with bullion and some are trying to pledge gold as a guarantee."
The amount of physical bullion purchased by private investors in the decade of the 2000's had ended at custodial accounts in various commercial banks. Some of these investors were discerning and suspicious enough to demand allocated accounts. Some were not. Unallocated gold can get pooled into a house custodial account with rights over custody being retained by the bank, not the investor. In this case, said investor owns not gold, but rather a bank liability payable in gold.
Unallocated gold in pooled accounts residing in a bank with growing funding stress makes for a rather easy liquidity target. The gold market offers depth in a broad range of currencies. Gold markets are also very well interconnected, between the physical market in London and various paper markets, particularly the CME in Chicago.
In the case of the large gold swap in 2010, the commercial banks accessed dollar liquidity "off-market" since the BIS simply held the bullion in its custody. Being accustomed to holding physical gold, it did have $23 billion, about 1,200 tons already on account, meant no additional hassle. The BIS surely incurred storage and administrative costs, but they would easily be absorbed by the interest rate the banks would pay on this collateralized loan (essentially the gold swap in this case amounted to a dollar denominated loan with gold bullion held as collateral by the BIS).
This "off-market" arrangement is the atypical feature here. A more typical leasing or swap arrangement finds bullion in the hands of a final counterparty that ends up disgorging the physical metal, such as a gold producer that may have previously sold short into the gold market to obtain favorable pricing. In this way, contrary to the BIS swap, gold leasing (collateralized lending with gold) ends up increasing the supply of gold for sale in the aggregate.
In the intervening steps, however, gold forward rates (GOFO) are supposed to be affected by this activity. GOFO is essentially the interest rate for a term loan collateralized by gold. It also sets the futures rates. In a situation where the demand for gold leasing rises, we would expect GOFO to rise as well. Where GOFO rises above LIBOR, the "leasing" rate is negative, indicating a relatively strong demand for gold collateral loans.
The reason GOFO is compared to LIBOR is an artifact of more normal conditions. It used to be that a gold owner would lease out gold (at GOFO) and obtain cash. The gold owner could then invest that cash in the unsecured lending markets (such as eurodollar) at LIBOR. If GOFO is less than LIBOR at whatever maturity, then the gold lessor earns the positive spread, called the gold lease rate.
However, in times of duress, the gold lease rate can and does turn negative. That means that gold owners need cash to the point where they will pay to lease out their gold - the opposite of what should happen in those normal conditions. Negative lease rates indicate high demand for cash using gold collateral.
In the past few years, LIBOR has become an increasingly meaningless number, and not just due to scandal. LIBOR is simply a survey of the largest banks recording what they would lend or borrow at in unsecured terms. It tells us little about any banks or institutions beyond them. Further, in today's interbank markets, there is little lending on an unsecured basis, meaning LIBOR tells us nothing about the cost of obtaining dollars. If banks want to obtain funding they better post some kind of usable collateral.
GOFO itself is a published rate of the London Bullion Market Association (LBMA) very similar in construction to LIBOR. The LBMA surveys its eight "contributors" and publishes rates based on what they would pay/receive for the tenor and product in question. Those contributors are: The Bank of Nova Scotia-ScotiaMocatta, Barclays Bank Plc, Deutsche Bank AG, HSBC Bank USA London Branch, Goldman Sachs, JP Morgan Chase Bank, Société Générale and UBS AG.
If LIBOR is not representative of the actual rate that the wider banking system actually pays for unsecured lending, then it could certainly follow that GOFO might be the same. After all, LIBOR has been far below 1% for several years yet European banks are seemingly bailed out or monetized by the ECB in almost perfect regularity. If the same is true for both these published measures, it brings up echoes of banking crises past.
One of the chief monetary "lessons" learned after voluminous scholarly study of the Federal Reserve and the depths of the Great Depression was how published and stated interest rates failed to capture the true picture of liquidity inside (and outside) the banking system. Published and benchmark interest rates were low (often record lows) yet these rates were not representative of reality beyond narrow constraints. Lending in most places was far more ad hoc, over-the-counter, rather than homogenous.
We saw a flare in gold leasing activity and this transformation of dollars for gold in late 2011. On December 7, 2011, the Financial Times reported that:
"Gold dealers said that banks - primarily based in France and Italy - had been actively lending gold in the market in exchange for dollars in the past week."
The rush to lend in gold terms had begun in September 2011 as sovereign difficulties again flared across Southern Europe, but it was impacting banks in dollar terms since US money market funds and other "liquidity providers" were trading out of those troubled assets. The rush out of Europe forced the Fed to offer dollar swap allotments that September, but that did not cure this dollar funding problem.
On November 30, 2011, the Fed finally relented to unlimited dollar swaps at a low premium (OIS + 50). But still banks were looking to gold leases. So much so that we have no idea at what rates these transactions were occurring. The same Financial Times article cited above quoted "traders" as indicating:
"...few, if any, banks were likely to receive the published rates since they have been skewed in recent months by a widespread reluctance among bullion banks to take gold for dollars."
The implication here is that "markets" had no reasonable idea how desperate for dollars some banks had become. It is no surprise in that context that the very day after the Financial Times published that article the ECB announced its massive lending facilities through the LTRO's. In conjunction with the Fed's swap lines, the two central banks, coordinating with other central banks, aimed to end the liquidity crisis through massive money stock means.
Before those measures, central banks had been active in leasing out gold to commercial banks, similar to the BIS arrangement from 2010. In January 2012, Thomson Reuters GFMS confirmed that central banks had increased the quantity of lent gold in 2011 for the first time since 2000. GFMS further said that commercial banks were swapping for gold with central banks in order to obtain dollars from the gold markets - thereby increasing selling pressure in the gold market itself. This undoubtedly occurred in conjunction with leasing both unallocated gold holdings and rehypothecation through futures markets.
The question now is whether or not we are seeing the same take place in 2013. Gold prices have acted exactly as they had in these episodes where gold leasing, or the gold collateralized lending, was ultra-prevalent. It certainly offers a compelling explanation, in my mind, as to why in the age of unprecedented central bank balance sheet expansion gold prices have faltered so badly. The data is pretty conclusive in these past instances, wherever collateralized lending in paper financial claims is stressed, gold appeals universally.
The implications are also straightforward - the banking system may yet again be short of usable collateral for repo funding no matter if published interest rates indicate otherwise. Since secured lending is now the backbone of the interbank marketplace, collateral availability is the prime factor in determining systemic liquidity in every major currency. Again, it is not a far leap to collateral shortage when so many central banks are undertaking measures that remove usable collateral. From the LTRO's at the ECB, to the Swiss National Bank converting its holdings of euros they obtained from maintaining the franc peg into German bunds, to the Federal Reserve removing both MBS and US Treasury collateral, to the Bank of Japan buying seemingly anything not nailed down. All of these measures that have been instituted to increase systemic "liquidity" do nothing but increase the static stock of money while reducing the agent for flow.
If we add in some sharp spikes in the demand for liquidity, perhaps countries and islands in the Mediterranean Sea and the growing propensity to confiscate deposits, the conditions for a gold lending crest have largely been met. Since LIBOR and GOFO are distorted and tainted indications, they do not give us a clean look inside these markets. In that respect, price might be the only true arbiter - particularly if leasing activity is enough to initiate secondary selling through increased hedging and technical positions.
Just a month ago, in response to the European Banking Authority's discussion paper on defining liquidity assets, Deutsche Bank's Global Head of Compliance Andrew Procter pointed out that gold-based liquidity is just as opaque as I have described here:
"As most gold transactions are over the counter (OTC), data is extremely difficult to access. This should not, however, stand in the way of the EBA conducting a thorough analysis of the liquidity value of gold. The World Gold Council will undoubtedly have a range of sources that it can make available. We would also highlight the efforts of the Bank for International Settlements (BIS), which has tried to compile derivative volumes globally, notably gold. Their studies (conducted triannually) offer detailed daily total amounts of OTC gold spot and swaps contracts etc."
Mr. Procter alludes to the "liquidity value of gold," something that I believe commercial banks all over the world have already assessed in their idiosyncratic struggles to navigate a changed and persistently challenging funding system. I also believe that it will become a phenomenon more familiar to gold investors, particularly since, if this historical pattern holds, the liquidity-driven slam in gold price is closely followed by frenzied "safety" buying. After all, growing illiquidity contrary to published perception is not an environment for elevating risk appetites.