Why and how negative GOFO really matters

Update: A few hours after this article was published 12/03/2014, LBMA announced its GOFO rates will stop 01/30/2015. As we wrote before, the world gold market is decoupling: the western part is fading quickly, but the eastern part is growing at the same speed. The next victim will be COMEX gold futures market. We can anticipate again a growing internationalized gold futures market in Shanghaï (SHFE International Board) at the same time. 

And remember the gold market is the unspoken foundation of International Monetary System. 


1- Accurate definitions 

GOFO (Gold Forward Offered) rates are the rates at which Market Making Members LEND gold to forward date on a swap against US dollars. 

GOFO depends also on the spread between the gold forward contract and the gold spot market. 

GOFO rates are submitted at 10:30am and fixed at 11am (London time), LIBOR rates are requested between 11.00am and 11.20am and fixed shortly thereafter. The LBMA since 7/1989 sets GOFO by polling its major bank dealers (Bank of Nova Scotia – Scotia Mocatta, Barclays, Deutsche Bank, HSBC, Goldman Sachs, JP Morgan Chase Bank, Société Générale and UBS, knowing that some of them like Deutsche Bank have recently shut down their gold desk, following the LIBOR and Gold fix scandals).[*] 

GOFO Rates means are given in % using 5 digits but their minimum absolute variation above zero is currently 1/600. 

Sober Look: “As banks and other investors realized that gold prices can indeed drop significantly, many decided to hedge their positions by putting on gold forwards to lock in the price. A gold forward involves a future sale of gold at a fixed price. Whoever sold them the forward becomes long gold for future delivery. The forward provider will hedge her position by borrowing gold (via a lease) and selling it into the spot market. Now the forward provider will receive gold in the future on the forward contract and can deliver it against her lease, and is therefore fully hedged.” 

fofoa: "As long as the GOFO rate is positive, the borrowing of dollars will cost you more debt service than borrowing gold. So it can be said that there is a bid from gold for dollars as long as the GOFO is positive! When it turns negative, it can be said there is NO bid from gold for dollars. 

Remember, a lease is where you "rent out" an asset to derive an income stream. And a swap (like GOFO) is where you need a loan, so you offer an asset as collateral and then YOU pay the income stream to someone else.” 

Source: LBMA

The forward contract is specific to LBM market. It is similar but not strictly equivalent to futures. In the GOFO market central banks are special warehousemen: they store only central banks gold, they can lease this gold to bullion banks, but usually do not sell or buy it in the open market (for instance because of CBGA agreement). In the strict perimeter of gold futures market warehousemen (bullion banks) store your gold but do not lease it. 

To calculate the daily derived gold lease rates (DGLR or usually called GLR), the GOFO rates means are subtracted from the corresponding values of the LIBOR (London Interbank Offered Rates) US dollar means: 

LIBOR minus GOFO = Derived Gold Lease Rate 

LBMA tables use the ‘LIBOR minus GOFO’ label. 

12 month GOFO data serie since 7/18/1989 (green); DGLR (blue); 
Source: Quandl.com ; updated daily [1]

GLOR (Gold Lease Offered Rate) is the daily mean of observed interest rates that central banks ask in order to lease the gold bullions they own. 

DGLR does not represent traded gold leasing like GLOR does. DGLR is a result of a simple arithmetic calculus. During normal market conditions, DGLR is a good enough proxy for GLOR. In this case LIBOR 3M should always be around 10bps (for storage etc) above GOFO 3M : DGLR = 10bps. 

Derived Gold Lease Rates from 1992 till 10/2013 (source

Derived Gold Lease Rates, 2 years moving data series 

We can observe: 1) 09/1999 was the biggest DGLR crisis till today. Explanations: 

In sept. 1999, 1M to 12M GOFO turned strongly negative but only for 2 days. Wednesday 29 Sept. 1999 is until now the only day where 12M GOFO was negative in LBMA history. It was the apex of the 1999 gold crisis : 
source: LBMA

We should note the 1999 gold crisis did not happened only for 2 days. The GOFO chart for this whole year depicts this clearly: observe how GOFO 1M has been divided by more than 2 in 97 and again in 98, before the 1999 crisis: 

GOFO rates, data series for Y1999; 
source: Conscience-sociale.org using LBMA data 

There are several major events to record here : 
  • the announcement of UK gold sale by BoE and Gordon Brown on 7 May 99 (i.e. : a crime of high treason against the english people; or watch this documentary about Brown's Bottom) 
  • the first gold auction on 6 July 1999; and others 16 auctions from July 1999 to March 2002 
  • and the historic first Washington Gold Agreement (CBGA) signed on Sunday 26 September 1999 
2) DGLR has been recorded with negative values during several months since 2009. Here is an explanation : 

Sober Look: “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. 

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.” 

The most important here, because GOFO rates are near zero or lower these days, is that LIBOR, GOFO, DGLR are each means of offered prices. They are not individual bid or ask prices. Rigorously speaking, no gold basis or gold backwardation status (i.e. negative basis and positive cobasis) can be calculated using only them. But they are still useful as we will see. 

2- On The Road Towards Permanent Gold Backwardation in the West

Fekete has explained since a long time why permanent gold backwardation is the worst case scenario. 

Weiner has explained how temporary gold backwardation is deepening towards permanent gold backwardation. 

An example of full backwardation (as defined by the distorted rates curve) occurred in March 1998 in the London silver market, when W. Buffet decided to take delivery of at least 110MM ounces. The following figure shows the SOFO term structure evolution during the months towards full backwardation. 

The silver forward rates are used here as a proxy for monitoring backwardation. It was more convenient and more readily available but also more approximate than the academic silver futures basis, until SOFO publication was stopped. 

fofoa has summarized how gold is withdrawn from the market during backwardation: “Only with a negative GOFO rate, you retain [at the end of the swap contract] control of the gold PLUS you receive the income stream coming in, so why instead would anyone LEASE their gold in this backwardation scenario? 

For that matter, why would you even SWAP your gold in this scenario? Unless, perhaps, you were desperate for dollars and you were swapping your gold with someone you knew for a fact would hold it safe for you and do nothing else with it.” 

But nowadays even central banks can make gold disappear without warning... 

Sandeep Jaitly has discussed the calculation of the gold basis using GOFO bid and offered prices in this essay

From our point of view, we can use publicly available GOFO as a trend or proxy signal for gold backwardation because : 

1) It seems that, on a very long timeframe, GOFO rates did not follow the same mostly always decreasing trend as did gold bases. But since 11/2011, they do. 

source and details (I've added the scale of the axes into the original chart) 

2) Since 1989, when GOFO 12M values become lower than 0,095 this has always been correlated with observed temporary backwardation for 1M contract (or worse) during the same week, as calculated using gold basis and cobasis. 

GOFO rates and "Paul's channel" with zoom till 12/02/2014; 
LBMA data compiled by conscience-sociale.org; 

This chart is updated frequently in our original article

What is remarkable is the way the purple line is drown (the bottom part of the channel), joining 4 key points of GOFO 12M curve:[2]
  • The 3 record lows : 05/22/2013, 05/06/2014, 12/01/2014 
  • And the value zero on 12/31/2014, the date of the end of G20 ultimatum to the US Congress to vote the negociated 2010 reform of IMF 
From this trend, we can anticipate when 12-month GOFO Rate will become negative (see channel above with yellow line and purple dash line). It will be the second time since the opening of western gold futures (or Gold forward) market. I've discussed this anticipation in this article and the first main consequences in this one

3- Beyond the screwed Gold Basis

a) The chartist way

Waiting for gold basis calculated using SGE futures international prices in renminbi instead of COMEX data, we can use the GOFO chart below to observe correlation between GOFO values, DGLR, geopolitical and monetary big events.

Source: conscience-sociale.org

This high definition picture allows to zoom (at the bottom) on events and key dates since 01/2010, together with some key DGLR values, record high or low. 

It is important to understand that because of ZIRP trap, the LIBOR rates cannot grow again above 0.5% like they did starting from 2004. Fed has lost this weapon for the next 2 decades (in US dollar life expectancy this means: forever), and cannot use it to sustain DGLR.

b) The basisGOFO way

What is the relation between Gold Basis and GOFO rates ? The answer is basisGOFO. 

Let’s follow Fekete’s questions

(1) What is the relation between LIBOR and LIBBR ? 

LIBBR is also called LIBID. 

LIBID (London Interbank Bid Rate) is the average of overnight interest rates for which banks accept deposits from others banks. It is published in the same time than LIBOR. 

For LIBID there is no formal fixing process, and it is supposed during normal market conditions that 

LIBID = (LIBOR – 0.125%) 

This means a difference of 1/8 base points. (source

(2) What is the relation between GLOR and GLBR 

GLBR (Gold Lease Bid Rate) is the daily average of interest rates that buyers want to pay to grab leased gold bullions; 

GLOR (Gold Lease Offered Rate) is the daily average of interest rates that market participants want to lease the gold bullions they own. 

During normal market conditions, these rates are observed with a 20 bps spread usually (source: LBMA/LPPM, page 21

This means GLOR = (GLBR – 0.20%) 

For GLOR and GLBR there is no formal fixing process. Instead, LBMA calculates daily a weighted average of GOFO rates. 

This calculation excludes the higher and the lower rates, as for LIBOR (this means they could be rigged the same way as LIBOR). 

Knowing the 20 bps spread discussed above, we can approximate the following values for GLBR and GLOR : 

GLBR = (average GOFO + 0.1%) 

GLOR = (average GOFO – 0.1%) 

“Average GOFO” is the currently published GOFO in this simplified model. 

We can then calculate the "refined GOFO" (I call it basisGOFO) during normal market conditions.

basisGOFO = LIBOR - GLBR = LIBOR - average GOFO – 0.1  

(3) What is the relation between GOFO and COGOFO? 

Knowing the LIBOR/LIBID spread above, we can calculate COGOFO (I call it cobasisGOFO) during normal market conditions.

cobasisGOFO = LIBID - GLOR 
cobasisGOFO = LIBOR – 0.025 - average GOFO 
cobasisGOFO = basisGOFO + 0.075 
cobasisGOFO = DGLR – 0.025  

(4) Can COGOFO go negative, and if so, what does it mean? 

Negative cobasisGOFO implies basisGOFO < -0.075 or DGLR < 0.025 during normal market conditions. 

Positive basisGOFO implies DGLR > 0.1 during normal market conditions. 

If you have access to LIBID, GLBR, GLOR exact historical data then you can determine much more precisely the trend of (co)basisGOFO during market crisis like nowadays, and explain movements between US dollar and gold, and the scarcity of each of them on the market. In short: at which extent each of them is used as a store of value. 


[*] This procedure might change in 2015. [update: it is now announced]

[1] The value used here for 07/31/2013 is questionable because Quandl.com has not refreshed the data they have downloaded from LBMA. 
As I wrote 10/28/2013: some values on the official GOFO page have interestingly been modified for the day 07/31/2013 since October 2013. July 31 was Comex August gold futures first notice day (see the gold futures calendar).
Previously we could read the following values for GOFO rates : 
31-Jul-13 -0.04833 -0.03333 -0.01167 0.06167 0.18000 
Now we can read : 
31-Jul-13 -0.04833 -0.04667 -0.03500 -0.01500 0.02000 
LIBOR data have also been modified accordingly for that day. 
Of course, LBMA does not support any commitment about the values displayed on its web site, and about any update which can occur at any position. We have never expected more from a financial services association.

[2] The record low on 31/07/2013 is not included but the reliability of this data is questionable (see note 1)


Le Référendum des Suisses sur l’Or : une Bonne Idée aux Conséquences Imprévues

Je publie ici la traduction en français que j'ai faite de l'article de Keith Weiner, "Swiss ‪Gold‬ Initiative: Good Idea With Unintended Consequences" publié cette semaine sur Forbes.com, avec l'autorisation de l'auteur.


Il se prépare en ce moment une initiative très intéressante en Suisse, sous la forme d’un référendum qui vise à imposer à la Banque Nationale Suisse (BNS) de maintenir 20 % de ses réserves en or. Les électeurs se prononceront le 30 novembre. Je ne cherche pas prédire le résultat du vote, mais je veux discuter ici de l'impact probable en cas de victoire du oui.

La plupart des analyses sur cette initiative porte sur le prix de l'or. Une prédiction typique est que ce dernier va monter, puisque l'achat d’or de la BNS excéderait l'offre. Cependant Mike Shedlock note que, "Quasiment tout l'or jamais extrait est disponible ...". En effet l'or est infiniment recyclable. La demande de la BNS serait faible par rapport aux stocks d'or présents à travers le monde, et c’est pourquoi le prix ne variera pas beaucoup. Shedlock ajoute: "Il est tout à fait possible que les achats de la BNS pourraient modifier de manière significative les perceptions ...". Je suis d'accord que le sentiment du marché est mûr pour basculer.

Le prix n’est pas une perspective très intéressante, sauf si vous êtes un trader sur le marché de l’or. Il est beaucoup plus important de noter que le référendum propose le premier changement monétaire positif depuis des décennies. Il réintroduit un lien entre l'or et l’activité bancaire, et impose un obstacle à la dévaluation monétaire. Pour cela, les Suisses sont des héros.

Il existe une faille critique dans notre système de devises flottantes. Chaque actif financier est une dette portée par quelqu'un. Lorsqu’une devise varie, elle crée des gagnants et des perdants. Les grands mouvements peuvent nuire aux banques dont les portefeuilles de prêts sont situés dans d’autres pays.

C’est pourquoi la BNS ne permet pas actuellement que l'euro chute en dessous de 1,2 francs. Pour maintenir cette parité, la banque centrale vend des francs et achète des euros. Il n'y a aucune limite à cette dévaluation du franc délibérée, qui dépouille les épargnants suisses, les investisseurs et les entreprises.

Les grands exportateurs comme Swatch et Nestlé peuvent avoir exercé des pressions pour un franc suisse plus faible, dans l'espoir de rendre leurs produits plus compétitifs, mais ce n’est qu’un aspect annexe. Le but réel de la dévaluation du franc est de protéger les banques suisses de la dévaluation de l'euro. Elles sont vulnérables, parce qu'elles accordent beaucoup de prêts hors du pays. Ils ont des actifs libellés en euro et les passifs libellés en franc suisse. Ils subissent des pertes lorsque l'euro chute, ou que le franc monte.

Deux exemples permettent d'illustrer le problème. Tout d'abord, imaginons que Jens en Allemagne emprunte un million d'euros auprès du Crédit Suisse. Alors que l'euro chute, Jens rembourse la banque en euro dont la valeur est de plus en plus petite. Sur les livres de compte en franc suisse du Crédit Suisse, la valeur du prêt chute comme une pierre. Jens est heureux, mais le Crédit Suisse ne l’est pas.

Ensuite, regardons Adriana en Italie qui emprunte aussi de l'argent, mais pas en euro. Elle obtient un million de francs suisses d’UBS. Alors que l'euro chute, Adriana le ressent comme un franc en hausse. Son paiement mensuel monte de plus en plus. UBS est heureux, au moins au début, parce que le prêt d’Adriana est en francs. Cependant Adriana se retrouve coincée. Quand elle ne peut plus rembourser, alors UBS devient aussi malheureux que le Crédit Suisse.

A chaque fois, le capital des banques suisses s’érode. Si l'euro chute suffisamment, alors les banques pourraient faire faillite. Seulement, ils savent où est la limite, et ce n’est probablement pas trop loin sous de la parité actuelle de 1,2 francs.

En cas de victoire du oui, les déposants des banques suisses ne vont pas ressentir la douleur causée par la devise, au premier abord. Ils sont heureux de posséder des francs suisses, surtout si le franc est à la hausse. Au lieu de cela, ils devraient se soucier des conséquences inattendues de la fin de l’ancrage à l'euro. Leur franc fort ne sera pas bienvenu dans le cas d'insolvabilité bancaire.

Malheureusement le régime du papier-monnaie impose un dilemme amer au peuple suisse. Ils n’ont que le choix entre des pertes lentes par la dévaluation ou des pertes totales par la faillite. Ils méritent une meilleure option, celle d’une feuille de route concrète vers l'étalon-or.

Il est formidable que le peuple suisse lutte pour se rapprocher de l'or. Je suis un ardent défenseur de l'étalon-or, et je tiens à applaudir mes amis suisses. Pourtant, je dois les mettre en garde aujourd'hui. Je me rends compte qu'ils ont dépensé beaucoup d'argent et de capital politique pour arriver jusqu'ici, mais je ne veux pas gagner cette bataille et perdre la guerre. Ils ont besoin d'une nouvelle initiative populaire, qui prenne en compte les prêts des banques libellés en euro.