Gold Market Sentiment Model – How it works


, , ,

I had a question the other day asking “Does term “sudden” in your chart relevant to second derivative? In any case I want to create graph of SD but I don’t know where I can get put/call data series on GLD and/or GDX/GDXJ. Is this publicly available information?”

I’ll answer the question above and a few others I’ve had in an attempt to describe the model without giving the whole model away.

Firstly a conceptual recap.  When investing in a market which has no cashflows and hence no valuation capability such as gold, I’ve found that you always need to know where you are in the sentiment cycle.  You don’t want to be the last one piling in at the top, or the last one giving up at the bottom.  Entering and exiting (or simply topping up or reducing exposure) should be done inversely to sentiment turning points.  That is, when sentiment is suddenly accelerating positively then more often than not, the last of the investors’ are piling in and you want to be selling into this.  Conversely when sentiment is suddenly crashing and all the investment flows are skewed to the downside, then investors are fearful and fully positioned for downside price movement – in those instances you want to be buying in the face of downside fear.

This approach works most, but not all, of the time.  Sometimes the downside fear, or upside enthusiasm, is vindicated in price action and prices keeping moving for longer than expected as external factors drive further price movements.  For example the Model has no idea that Janet Yellen was about to speak last night, and hint that U.S. rates would stay anchored for longer than the market was pricing in.  However, using this model serves to minimise the damage when you are over-exposed in a falling market and maximise the reward when you are fully exposed in a rising market.  Over-time you can start to out-perform by ignoring the market noise and almost holding positions inversely to the view of the masses.

Back to the question, the answer is yes the model is all about the second derivative.  Or put another way, it’s about measuring and monitoring turning points in sentiment over time, as opposed to basing decisions on absolute sentiment levels.  The Model tends to change from “buy” to “sell” only about 6 or 7 times a year, and lasts anywhere from a few weeks to a few months at a time, before sentiment swings around again and the model switches from buy to sell, or sell to buy.

For the model to switch the “sentiment score” (measured as aggregate Put Premium divided by aggregate Call Premium) needs to change by a pre-specified % amount over a pre-specified number of trading days.  These pre-specified model inputs are what drives the success of the model and have been hard coded into the model after years of testing.

For input data, I use the entire COMEX option premium array for the most liquid maturity futures contract – eg Gold Options on April 2015 COMEX futures.  I wouldn’t use GDX or GDXJ as these are equity indices.  GLD is not a bad proxy, but options data is less liquid.  The other benefit of using COMEX and other exchanges is that from experience I know that the gold Over-the-Counter (OTC) market uses the exchanges to arbitrage away any OTC action.  Therefore you are picking up the sentiment of the gold OTC market which is the biggest market in the world for gold trading.

Finally I get my data from my Bloomberg subscription.  However in my early days I remember going onto exchange web-sites and simply typing down each day’s closing prices for each outstanding option contract.  I did this everyday for a year or two, before launching my Fund inside a hedge fund group and latching onto the time saving Bloomberg access.  However at $50,000 a year it’s not available to everyone, so try the NYSE website for GLD or COMEX website for Gold Futures.

Gold Market Sentiment Model


, , , , ,

I’m considering posting my gold market sentiment model results on the blog later this month.

Over the next several posts I’ll aim to introduce the model I have been using to manage gold market investments over the past 8 years.  I used the model systematically every day for 4 years running a $100m gold market hedge fund and then another 4 years running a $160m UCITS gold fund.

I developed the model in the early 2000’s when I noticed that central banks’ and gold miners’ were exiting the gold options market, and it was being taken over by the investor community.  Finally a real investor driven window into investor gold price expectations was coming available.  The options market has been used for years in equities markets to measure investor sentiment by comparing the demand for upside exposure (call market prices and activity) to downside exposure (put market prices and activity).

I collected the put and call premiums on all available strike prices on gold exchanges from 2002 to 2006 and plotted the resultant aggregate ratio over time.  What I noticed was that prior to big rallies in the gold price, sentiment actually deteriorated (i.e. demand for calls dried up and demand for puts was accelerating).  Conversely prior to gold price falls (back then it went up the escalator and down the lift regularly) sentiment was often improving (i.e. call demand was increasing more than put demand).

You can see that sentiment was a contrarian indicator, i.e. investors as a whole were getting the short term price direction completely wrong.  When they were overly bullish (and Bloomberg was full of positive stories about the yellow metal) the available market was fully invested – so any bad news was hit hard and investors headed for the exit.  When they were overly bearish the available market was under-invested – so any good news was met by above average buying and the gold price tore higher.

I set the parameters in 2006 and they haven’t changed since.  A back-tested accuracy rate of 70%, was followed by 70% and 65% accuracy in the hedge fund and the UCITS fund.


The gold market is unique in many ways from other financial markets.  Its performance is often uncorrelated (and even negatively correlated) to the vast majority of asset classes.  In addition the return drivers for gold prices are often unclear and subject to frequent displacement.  For months it will be driven by a negative correlation to the US dollar, only to suddenly ignore US dollar movements and become a function of real yields, or physical bar demand or an escalation in geopolitical tensions.

The difficulty in modelling the price of gold stems from another of its unique qualities; it does not intrinsically generate a return.  Cash-flow based valuation techniques such as Net Present Value or Dividend Growth Models cannot be used to generate a reference point for gold’s intrinsic value.  The same could be said for commodities in general, however again gold is unique in that its pool of above ground supply is never extinguished; unlike energy or other metal products whose supply is constantly being consumed and future demand has to be met by future production.  In the case of gold every ounce of gold ever mined still exists.

The lack of a verifiable source of valuation leaves the determination of gold’s current value firmly in the hand of its majority holders (central banks), gold miners and most importantly these days – investors.

Since around 2000 both central banks and gold miners have reduced their influence over gold prices dramatically.  During the 1990’s central banks were very active in the gold market, their activities ranging from pre-announced selling programmes to premium generating activity such as the granting of call options and leasing their gold reserves to gold miners.  The gold miners were keen to borrow as this allowed them to sell gold in the spot market (years ahead of actual mine production) and deal in the gold options market in the now under-active practice of “gold price hedging”.

Since 2000 these activities have all but dried up, and now gold price movements are driven primarily by investment demand, albeit supported by physical demand from the jewellery and to a lesser extent the industrial market.  The options market is dominated by investor behaviour where investor sentiment can now be meaningfully tracked.

As often happens when the market gets too bullish or too bearish, conditions become ripe for a reversal. Unfortunately, the crowd is too caught up in the feeding frenzy to notice. When most of the potential buyers are “in” the market, we typically have a situation where the potential for new buyers hits a limit; meanwhile, we have lots of potential sellers ready to step up and take profit or simply exit the market because their views have changed. The put/call ratio is one of the best measures we have when we are in these oversold (too bearish) or overbought (too bullish) zones.

The model excerpt graph below shows the warning signals, when market reversals were nearby.

GMSM Aug 2011

Source: Gold Market Macro

In the next few posts I’ll elaborate on the model and start to show the performance over time to help you manage your gold market portfolios.

Central Banks can’t stimulate demand and will ruin their currencies trying


, , ,

Yet another Central Bank (this time Australia’s RBA) has tried reducing the cost of borrowing in the economy that they imagine they are commanding.  The theory sounds sensible enough – reduce the base-cost of borrowing and pressure all lenders to pass this reduction through to grateful borrowers – who in turn borrow more or use their increased disposable income to make investments, hire employees or increase their retail expenditure.  The problem since the confidence implosion of 2008, is that western economies (private and particularly the public sector) are not interested in borrowing more – in fact they’re de-leveraging their balance sheets wherever possible.  Therefore lower and lower rates are not stimulating end-demand.

So will the RBA turn to other monetary policy techniques such as Quantitative Easing when they reach the lower bound of interest rates.  That is, when the “price” of money doesn’t help, will they try the “amount” of money next like the BOJ, the FED and the ECB before them?

This will be a dilemma that all central bankers are debating at the moment.  They are by now aware that the most fundamental shortcoming of QE – or, in fact, of using monetary policy in general to combat a recession – is that it only “works” if it somehow induces the private sector to spend more out of current income.  QE is simply swapping longer dated sovereign bonds for shorter dated reserves – none of this stimulates end consumer demand (except slightly through the temporary wealth effect of rising stock markets).  Unfortunately for rate setting central banks the direct route (fiscal policy) is out of their sphere of influence.  Given the much-needed deleveraging by the private sector, the best technique would be to target growth in after-tax incomes and job creation through appropriate and sufficiently large fiscal actions. Unfortunately, stimulus efforts to-date have not met these criteria, and the prevailing obsession with “strong public finances”, which dominates policy discourse in the West will ensure that their economies remains stagnant.

All this points to continued loose but ineffective monetary policies for years to come.  Insofar as gold trades as an alternative currency which is not subject to wave after wave of unsuccessful monetary policy experiments, it will start to trade at a significant premium to fiat currency alternatives.  Physical holdings of gold will surely outperform holdings of fiat alternatives as a way of handing down wealth from one generation to the next.  This is being increasingly appreciated in China, India and Russia at the moment, and in parts of Switzerland, but for rest of the world they continue to hold onto the coat tails of well meaning but ultimately ineffective central banking policies to sustain their wealth.

Low gold prices are a blessing in disguise for physical gold accumulators


, , , , ,

Gold has begun an irreversible movement into strong hands that will hold on to it, and will not relinquish it, even in the face of steeply rising prices. Fallen prices since September 2011 have been a blessing in disguise for physical gold accumulators. A lower price for paper gold has made it easier for participants to demand delivery on maturing futures contracts; exchange balances are being siphoned; ETF balances are being raided, recast and sent to Hong Kong as kilo bars before disappearing into China forever; central banks are scrambling to repatriate gold back within their physical control; mine supply is being earmarked well before it reaches the surface; and Gresham’s law together with gold’s Veblen good characteristics means higher prices will not necessarily tempt the general public to feed the scrap supply monster.

As these, and other similar trends, converge and continue apace, the price of the nearby futures contract will drop to previously unimaginable depths, relative to the cash price, making backwardation worse. Ultimately this will make backwardation irreversible and only those that have acted in advance will be positioned to transition well to Bretton Woods II.

Permanent gold backwardation

As we reach the end of the year the outlook for gold in 2015 remains uncertain. Gold has been in backwardation 40% of the time during 2014 and remains there today. Given the uncharted financial territory in which the global economy currently finds itself it seems likely we will continue to witness the same market participants continue their accumulation of physical gold, creating further physical tightness in the gold market in 2015.

141231 GOFO YTD

We wish all our readers a happy and prosperous 2015!

Physical tightness extends further along the curve


, , , , ,

The best indicator that the physical market is becoming stressed will be seen in the forward curve for gold.  Over the past week GOFO rates published by the LBMA have all increased indicating fresh supply of gold available on loan, or a reduction in demand to borrow gold.  This has normalised the gold curve out to 12 months.

I read opinion suggesting that this normalisation can’t be trusted because member contributors mis-reported their quotes to avoid highlighting the problem.  I think gold commentators need to avoid calling foul every time the price or other market variable goes against their prediction or desired direction.

Instead by digging a little deeper we can see that this normalisation doesn’t tell us about the rest of the forward curve – beyond 12 months.

As a counterparty to a 5 year lease rate swap agreement with a major bullion bank, I see daily movements in this part of the curve.  Interestingly, even though short dated GOFO has normalised (and therefore short term gold lease rates have fallen) over the past week, the valuation of my lease rate swap agreement (to pay fixed lease rates and receive floating lease rates every 3 months for 5 years) has increased.  This means that over the past week, the market’s view on “average lease rates over the next 5 years” has actually increased – despite the decease in lease rates from 1 month to 12 months.

This means the market’s demand to borrow gold has simply moved further out along the forward curve.  In other words, institutional participants are expecting a tight gold market to persist for years to come, and are taking action to borrow gold now before it’s too late.

Gold backwardation is no “storm in a teacup”


, , , ,

A recent post from The Speculative Investor argues that the negative GOFO rate has almost nothing to do with gold supply/demand:

The “gold backwardation” (a.k.a. negative GOFO) storm in a teacup

The writer states that “a negative GOFO effectively just means that it costs more for a major bank to borrow gold than to borrow US dollars” and argues that this development is neither strange nor important.  However, gold backwardation says something about fiat currency demand versus gold money demand.  We are living in a post gold standard monetary system – trillions of US dollars are pledged, stored, invested and lent without any link to a gold system.  Gold has been relegated to a pure commodity, and a commodity whose supply increases every day and never falls – every ounce ever produced continues to exist.  In that environment the demand to borrow gold should be minimal at best.  On the other hand fiat currency is in demand globally and at all times, and as Marshall Auerback explains so well, will always be in demand so long as governments dictate that taxes are to be paid in fiat currency.

Not only that but borrowing US dollars for a major bank has no net cost – they can borrow US dollars at LIBOR and lend it at LIBOR.  Contrast that to borrowing gold on swap where they have to give up their US dollar yield (USD posted on swap) and now they have to pay for storage of gold.  So all things being equal the modern day banker should much rather borrow US dollars than gold.  The fact that demand for borrowing gold presently exceeds the demand for borrowing US dollars (GLR > LIBOR) means there is some other reason to borrow spot gold which can only be called a convenience yield (i.e. some motivation (fear) to hold, finance and store gold now, rather than rely on a claim for future delivery). In oil that can make sense if a supply disruption is feared (geopolitical unrest) because your future delivery may never arrive and you have on-going needs to use your oil, but with gold future delivery should never be in question UNLESS you start to believe gold is going into hiding and won’t be for sale at any price in the future.  In other words a severe backwardation can mean that demand for fiat currency is plunging and gold is being re-monetised unofficially by large swathes of the global economy who are buying with the intention of never selling (sounds like the Swiss gold referendum which was quashed by orthodox central bankers).  We’re not there yet in the West, but the discussion is still worth having.

A Golden Bank Run


, , , ,

Bank runs are terrifying events that have been part of every fractional reserve system in history.  One moment your capital is simply on-demand, one simple withdrawal request away, and the next moment you’re facing either a hair-cut or the full force of a complete default.  Negative GOFO prints are accelerating (1 month GOFO fell to -0.6% yesterday), and I think these are the best evidence that a “golden bank run” is underway.

From here, like every bank run, it’s all about confidence.  1.  Confidence from the gold depositor (think European wealthy family offices, Arabic oil states, Russian Oligarchs, etc) that the bullion bank custodian can return the gold on demand. 2. Confidence from the bullion bank that the gold they on-lent to the mining companies or short side speculators, will be mined and returned. 3. And Confidence from the generalists that if they ever decide to buy gold that it will be available on-demand.

I’ve been studying lease rates and GOFO for years, and I think real panic is brewing within certain gold depositors.  Whether it’s been all the talk about gold repatriation, high and rising physical premiums in China and India, falling registered COMEX gold levels, technical default by several European banks who elected to close gold accounts and pay-out in cash only, retail rationing at global mints or more likely first hand experience of a bullion bank pushing back on requests to return their gold deposit.

This is one reason why I can’t understand why wealthy investors would turn themselves into bank depositors – why would they trust bullion banks to store and manage their gold for them when the whole case for gold holdings is to completely eradicate any form of third party credit risk.  It’s always the depositors that fare worst and have the most to lose.

From the bullion banks perspective (remember they are the ones setting the GOFO rate) negative GOFO means instead of accepting gold deposits, posting $ in return and charging the depositor the GOFO “yield” they are now so desperate for gold deposits that they are advertising to accept gold, post $ in return PLUS pay for the privilege (0.6% annualised for rolling 1 month agreements).

Think about it again.  The bullion bank is accepting an asset that has no intrinsic return, has expenses to insure and store, and in return is posting $ that it could otherwise invest in the markets, $ that it otherwise has to raise via it’s individual cost of capital.  For this exchange, they themselves are coming up with a negative charge!

Now the question to ask is why?  Like any fractional bank, the bullion bank is mismatched between deposits and loans.  What is happening is that they are being hit from both sides.  Depositors are lining up demanding return of their bullion and at the same time their borrowers (mining companies) are struggling with their own default probabilities (see CDS on gold miners – its on the rise as gold prices near the cost of production).

Also bullion banks do not simply keep the gold on account.  They mobilise it in their own investment books, they rehypothecate it, they register it on exchanges for futures contracts delivery, in summary they can’t always get it back.

Being hit from both sides like this makes a default progressively likely.  The most aggressive gold depositors will be best placed, as in any bank run.  First in, best dressed.  As the panic spreads gold’s physical premium (convenience yield) will shoot higher, and overnight gold will go bid only.  I would recommend you get your physical requirements in order well before that moment.

Part II – Eastern Physical Demand meets Western Paper Supply


, , , , , , ,

How do we square the poor price performance of gold these past few years with an increasing physical demand picture? As the writer points out in this linked article, if we compare apples with apples and look at Shanghai Gold Exchange (SGE) withdrawals as a proxy for end-demand you can see that Chinese demand is rising sharply.

SGE figures – do they really equate to Chinese gold demand?

Also he says “Thus global gold demand remains at a huge level, probably well in excess of newly mined production plus scrap which makes recent gold price performance anomalous to say the least, giving ever more ammunition to those who firmly believe the price is being suppressed for whatever reason.”

That’s the point we were making yesterday in our investment meeting – if demand is so high, why is price performance so weak? If it’s not suppression (which I don’t discount completely) then it comes back to the irrelevance of annual increments of supply and demand for a monetary metal like gold (where above ground inventory is 50 times annual production). For combustible commodities like energy products, annual (and even monthly) ebbs and flows in demand and supply have a disproportionately large impact on price because the above ground inventory is so small compared to demand levels and the threat of commodity shortages are real.

For gold, however, all the gold ever mined is still available. Now this irrelevance of incremental supply and demand could be changing I think. The crux moment for gold will be when (and if) all the gold ever mined becomes spoken for, or when it’s held so closely to users’ chests that it’s effectively removed from the market – and in an instant gold changes from being priced with no “convenience yield” to having an infinite convenience yield. At that moment, spot gold will trade at a premium to future claims on gold (paper gold) and the spot price could spike astronomically higher. Then even the smallest release of physical gold onto the market will be snapped up by those who need it most (think the shorts who will be called to deliver on their promise to sell at much lower prices!).

There may be plenty of investors who find themselves short gold in this environment, not just those who have sold futures contracts or sold call options. There will also be a sizeable group who previously believed that they had gold, but when the dust settled found out that their claim on gold was no longer deliverable. These participants will then have to re-enter the market to get the exposure they thought they had. The Dutch central banks has made sure this past couple of months that they will not be one of those parties. Well done them for securing their gold reserves more closely (although not completely yet) to home soil. It makes the German effort to have their gold reserves stored on-shore look paltry by comparison.

Let’s hope the Swiss deliver some fireworks this weekend, and 1 month GOFO can keep heading further into negative territory. (last -0.33% p.a.)


Eastern physical demand versus Western financial supply – who will win out?


, , , , , ,

1 month GOFO moved from -0.08% to -0.185% yesterday, and 2 month GOFO recorded -0.145%, which is the lowest level since the 1999 Washington Accord gold supply panic.
This negative gold forward rate is a true backwardation of gold prices.  It means participants can sell physical gold now and instantaneously buy it for delivery in 2 months time at a 0.145% discount.  Not only that, but the entrepreneurial arbitrageur can also invest the proceeds from the spot sale in risk-free securities over the period.  In two months time, they will have their gold back, have banked 14.5 basis points profit, have banked the risk free security return and will have save on 2 months storage and insurance of the physical gold. 
This is why backwardation of gold is so fascinating, it just should not happen.  The arbitrageur may not be so smug if the future delivery never occurs (in trying to make 0.145% he has lost 100% of his gold), and that is the risk which backwardation effectively prices in. 
Why it is happening now is due to one of two things.  Either it means there is the sniff of panic around physical gold buyers, in so far as the gold inventory they require over the next two months may not materialise.  Or it means the demand to short the gold market has become so high that financial investors are prepared to borrow gold (to short it) at rates higher than borrowing unsecured funds such as Libor.
So in other words we have a situation where physical gold demand is rising but financial demand is weak, in fact financial demand has become so weak it’s turned into financial supply.  Which side do you think is taking more risk in this scenario – the buyers of physical gold who use their savings to accumulate gold they never intend to sell or the sellers of financial gold who use borrowed money to short gold they will one day have to buy back! 
Looking in more detail at the first reason, imagine you are the Perth mint or a Chinese or Indian bank or gold merchant with a steady stream of physical gold buyers on your customer list.  Then imagine that prices have fallen dramatically and your buyers are calling you all hoping to take advantage of the lower prices.  Suddenly you order books are filling up fast which is great, but what if your smelting contacts advise you there’s a problem with casting the required volume of bars and you’ll only get a portion of your order unless you pay a premium for spot delivery.  This premium amounts to a “convenience yield” which is more common amongst commodities that can “run out” if production stops like the oil products.  In other words you can have more, but you’ll have to pay more for today’s gold, than if you’re prepared to wait and lock in a price for future (uncertain) delivery.  If a real global run on physical gold occurs spot prices will sky-rocket compared to contractual claims on future delivery.  Products like Gold Backed ETFs and financial futures prices will trade at an even greater discount to spot gold delivery.  Spot gold physical markets will trade at a premium and gold backwardation could become the norm.  In that scenario anyone prepare to lend gold to a starved physical market can almost name their yield (lease rate).  Who will be saying that gold has “no return” in that environment?
I’ll elaborate on the second reason in another post.   

A question of trust


, , , , , ,

If “trust” was a tradable commodity its price would be sky high right now; demand is high and yet supply is extremely low and existing stocks are perishing. Think of the events of the recent past where trust has been lost – the Worldcom/Enron accounting scandals, the interbank market meltdown in 2008, the Fed/BoE/ECB/BOJ experimental monetisation of government debt, the Occupy Wall Street movement, uprisings across the Middle East and the Snowdon revelations evidencing mass international spying.

It follows then that the Pew Research Center’s survey of Americans recently showed that one age group (18 to 33 years-olds) recorded the lowest ever result (19%) for the trust question “would you say that, generally speaking, most people can be trusted?” And according to Dr. Jane Holl Lute, president and CEO of the Council on Cyber security and a former deputy secretary at the department of homeland security, with regards trust “the trend of decay is the near total collapse of public trust in public institutions, and it’s true globally. So what we are seeing fundamentally is the rise of the human being taking matters into their own hands”.

Washington Post: Millennials’ lack of trust part of global trend

As stinging as that statement is, it does establish “trust” is near record low levels; so shouldn’t gold prices be near record highs? Instead gold spot prices, in US dollar terms, are languishing around 30% below their peak in 2011. On the surface that seems incongruous and has many investors puzzled.

A recent article by Steve Ellis looks deeper and establishes that we are witnessing record levels in the gold market (just on another metric). The gold forward curve (the series of tradable prices spanning out at fixed future dates) has become downward sloping (or “backwardated”) for the longest consecutive period in gold’s history. The piece discusses whether this sporadic backwardation is about to give way to permanent backwardation and what that means for financial markets.

Permanent gold backwardation