Gold is Money (Sort Of)
Few markets are as widely misunderstood and subject to as many misconceptions as the gold market. Many of these misconceptions stem from gold’s dual characteristics as a commodity and money. Is it actually correct to claim that “gold is money”? After all, it is not used as official money anywhere and barring isolated instances of payments made from digital gold accounts, it is unlikely that one will ever make payments in gold these days.
In addition to this, central banks have been intent on “demonetizing” gold, and many of the biggest central bank holders of gold (except the US) have unloaded their gold reserves for years, ostensibly in order to earn the higher returns provided by bonds. It has always struck us as odd that they would be selling gold for this reason. The primary purpose of central bank reserves should not be to produce a return, but to enhance confidence in the currency. However, central banks are remitting their profits to governments, which seems likely to have been a major reason for this repurposing of reserves.
In spite of all this, we still know that gold is money in a sense, because it trades in the market as if it were money. This is to say, it is not really money at the present time, since we cannot use it to effect payments for goods and services in the normal course of business, but it nevertheless retains its monetary character. Surely, the markets are implying as much: If gold’s price were determined by fabrication demand alone (jewelry and industrial demand), it could not possibly trade at a price of $1,250/oz.
In the late 1960s and early 1970s, many mainstream economists advocated the repudiation of the gold exchange standard and applauded Nixon’s decision to default on the US government’s gold obligations (“temporarily”, as he stressed at the time). It was a widely held view among them that following the official “demonetization” of gold, its price would decline from the then official fixed exchange rate between $35 and $42/oz. (the $35/oz. price was decreed by FDR way back in the 1930s on a whim) to about $5 to $6, as its price would begin to solely reflect its use value.
They turned out to be mistaken, as gold did pretty much the opposite. Instead of declining to $5, it eventually soared to $850 by early 1980. The worldview of supporters of central economic planning naively presupposes that government can magically suspend economic laws. Given that Keynesians and monetarists alike advocate positivism (empiricism) in economic science, one must actually wonder if they even admit to the existence of economic laws. If asked, they will certainly reply that they do believe in economic laws, but this seems to be at odds with their positivist approach, which is not suitable to discovering and proving the validity of such laws (the situation is different in the natural sciences). However, we digress. The important point with regard to gold is only this: It turned out that it was actually not possible to truly “demonetize” gold by government decree. The markets decided otherwise.
The world’s estimated total stock of gold and its annual rate of growth.
Gold Supply and Demand
Many gold analysts, from the mainstream to “fringe” groups such as the Gold Anti-Trust Action Committee (GATA), are in the habit of making gold price forecasts by adding up annual fabrication demand, central bank purchases, de-hedging demand by mining companies and (so-called “implied”) investment demand, and contrasting the resulting total with annual supply (mine supply, central bank sales, disinvestment and scrap). In short, they analyze the gold market in exactly the same manner as they would e.g. analyze the copper market.
It should be immediately obvious that this cannot be the correct approach. After all, nearly the entire amount of gold ever mined (approximately 170,000-175,000 tons) still exists. The total potential supply of gold is therefore many times greater than the amount of gold produced every year (approximately 3,000 tons currently). There is enough gold held in inventory to theoretically supply fabrication demand for decades.
Thus, it makes no sense to apply traditional commodity supply/demand analysis to the gold market. There is a big difference between commodities that are effectively used up (aside from scrap that is returning to the market in some cases) and a commodity the indestructibility and durability of which inter alia made it an ideal “money commodity” in the first place. Annual flows from mine production, the annual demand from industry and warehouse statistics are certainly helpful in gauging the copper market’s likely direction; the same data are essentially useless in analyzing the gold market.
Jewelry Demand vs. Monetary Demand
One can further illustrate gold’s unique nature by comparing gold prices to jewelry demand. If record fabrication demand for gold was good for the price of gold, then a historic high in jewelry demand should in theory coincide with a very high gold price. However, record high jewelry demand in 1999-2000 in actual fact coincided with a 20-year bear market low in the gold price – the exact opposite of what traditional commodity supply/demand analysis would suggest.
We can conclude from this that there must be a source of gold demand that is of far greater importance than jewelry and industrial demand, and that this demand constitutes the true driver of gold’s valuation in terms of fiat money. Indeed, there is; this demand component is called monetary demand, or investment demand. Gold’s monetary supply and demand situation to a large extent reflects the degree of reluctance of current owners of gold to part with their gold at prevailing prices. Given that some 175,000 tons are already held in stock by various market participants, their so-called “reservation demand” is in fact the most important factor in gold price formation (for another look at how the gold price is formed and the importance of reservation demand, see this excellent article by Robert Blumen that we published in 2011: “What Determines the Price of Gold?“).
The markets treat gold as if it were a currency rather than a commodity. It often keys off other currency cross rates, such as dollar/euro and has a strong tendency to do the opposite of what the supply/demand analysis typically employed by many mainstream sources (including the World Gold Council, which should know better) would suggest. A rising gold price as a rule begets a noticeable decline in jewelry demand, i.e., this demand component is subject to price elasticity. By contrast, rising prices actually tend to stoke investment demand to some extent, just as a developing uptrend in the stock market tends to have a positive effect on market psychology.
Jewelry demand and the gold price; ignore the “total demand” line – it is meaningless.
The above chart shows that the record high in jewelry demand coincided with a 20-year low in the gold price. Obviously, jewelry demand has not been what has driven the price of gold higher since then. In fact, jewelry demand for gold has been declining, resp. flat, since its 1999 peak.
The chart also shows that concurrently while the price of gold fell from 1980 to 1999, fabrication demand steadily increased. This is further proof that fabrication demand is not an important driver of the gold price – on the contrary, it is exactly the other way around. As an aside to this, it should be noted that some jewelry demand, especially in India, in reality represents monetary/investment demand in disguise.
During gold’s recent cyclical bear market, one of the arguments trotted out to “explain” the fall in the gold price and predict further declines, was that the gold holdings of ETFs such as the SPDR Gold Trust ETF (NYSEARCA:GLD) decreased (and further decreases in their holdings were forecast). In 2013, GLD shed about 500 tons of gold, some 37% of its entire hoard at the beginning of the year. However, this tonnage represents just 0.28% of the total supply of gold. To see the absurdity of the argument that this is relevant for the gold price, one only needs to think about it in terms of the dollar (or any other fiat currency). If a currency analyst were to state: “The dollar will decline next year because the supply of dollars has increased by 0.28%”, people would look at their calendars to check if it was April Fool’s day.
In fact, changes in gold ETF holdings are a consequence rather than a cause of gold price movements. When gold buyers are more eager than sellers or vice versa, GLD will often trade at a premium or discount to its net asset value during the trading day. This is an incentive for authorized participants to engage in arbitrage; if GLD trades at a premium, they will create new baskets of GLD shares, while concurrently buying the physical gold required to back these new shares. They then sell the new shares and pocket an arbitrage profit in the process. The opposite happens when GLD shares trade at a discount to their net asset value; then it makes sense to retire baskets of GLD shares and sell the underlying gold.
GLD’s price and tons of gold held in trust.
Nearly all major brokerage houses and banks have been prone to attaching importance to jewelry demand and similar demand components in gold price forecasts in the past (there has been some improvement on this front in recent years, as a number analysts have begun to shift their focus to the truly important price drivers). At times, this has led even some very astute money managers astray. For instance, legendary value investor Jean-Marie Eveillard admitted to this error in an interview with Fortune in 2007 (unfortunately, Fortune has in the meantime scrubbed the interview from its servers, so we can no longer provide a link to it):
“When we started our gold fund in 1993 – which proved to be six or seven years too soon – I mistakenly thought that my downside was protected by the fact that jewelry demand was fairly vibrant. But I was wrong. I think gold moves up and down based on investment demand mostly.”
We would go even further and replace “mostly” with “only”. The WGC (World Gold Council) tries to gauge “implied investment demand” or “implied disinvestment” retroactively at the end of every year, by adding up the known annual additions to supply (from mining, scrap, central bank selling and hedging) and contrasting them with annual fabrication demand and other measurable sources of demand (CB purchases, de-hedging). It reckons that the difference must represent “investment demand”, resp. “disinvestment” (demand from gold ETFs figures in the calculations of investment demand as well these days).
In the London market alone, an amount equal to annual mine supply is changing hands every three to four trading days. We can be reasonably sure that this is not from jewelers trading gold back and forth. However, as noted above, investment demand is also – even mainly – expressing itself by the reluctance of current gold holders to sell at a given price. Their reservation demand cannot be measured, and actual investment demand is therefore also not directly measurable. All we can state is that total demand must equal total supply.
Global annual gold production from mines, 1900-today.
The above chart of mine production depicts another peculiarity of the gold market. As the price of gold rises, mine production actually tends to flatten out and decline. There are a number of reasons for this seemingly unusual response to rising prices.
- During low gold price environments, miners are forced to “high grade” their production (i.e., they will focus on mining the higher grade portions of their ore bodies). Once prices rises, they tend to shift mining activities to the lower grade portions of ore bodies, which have previously not been economically viable. This lengthens mine life, and allows more revenue to be wrung from existing mines. However, it is difficult to increase the production capacity of mines that have been in “high-grading mode” for a number of years.
- During periods of low gold prices, exploration and development spending falls. Once prices rise, very few new mines are set to open and take up the slack from depleted mines. It can take up to ten years and sometimes even longer from the discovery of an economic ore body to the point when mining can begin. In addition to this, the previous bear market will still weigh on the minds of most mine managers in the early stages of a nascent bull market. Initially they won’t trust the new trend and as a result will be slow to take advantage of it.
- Conversely, by the time a gold bull market is over, many of the new mine developments that have been initiated while it was underway will finally come online. Thus production will increase just as prices are embarking on a downtrend. Moreover, since developing new mines is very costly and often involves the assumption of sizable debts, the initial reaction to declining prices will be to do everything to increase rather than decrease production, in order to obtain the revenues required to service these debts.
In one sense, gold mining companies are, however, in a very fortunate position. Whether they increase or decrease production is pretty much irrelevant for the gold price. The same principle that we have already mentioned in the context of jewelry demand and ETF demand applies to mine production as well. If annual gold production jumps by 20% from 2,500 to 3,000 tons, this will also represent a mere 0.28% increase in the total supply of gold – the effect on prices will hardly be noticeable.
What Drives Monetary Demand for Gold?
Let us look back at how gold became money in the first place. Initially, there was demand for gold based on its usefulness for creating ornamentation and jewelry, i.e., before it became a medium of exchange, there was demand for it due to its use value. In a process of trial and error, it turned out that was very useful in barter transactions, this is to say, its marketability proved to be very high. The more it was employed as a medium of exchange, the more its marketability grew. Due to this growing marketability, it increasingly replaced competing media of exchange and its exchange value rather than its use value became the major determinant of its valuation.
Gold’s durability, divisibility, fungibility and easy portability all contributed to making it useful as money. Lastly, the fact that its supply is unlikely to suffer sudden increases, regardless of the wishes and actions of the money issuing authorities, has made it extremely useful as a long-term store of value as well.
There exists just one historical exception to the idea that the stock of gold isn’t going to be subject to sudden large increases. When Spain began to import large amounts of gold to Europe from the New World in the 17th century, it increased Europe’s gold supply significantly in a very short time. This led to considerable price inflation, as gold’s purchasing power declined. However, nowadays the annual supply from mining on average represents only about 1.5% of the total stock of gold and it is highly unlikely that any large deviations from this percentage will occur in the future.
Gold price, 1968 to 2015.
The fact that the government and the banking system can’t create gold out of thin air is obviously a very important factor driving monetary demand for gold. In the modern-day fiat money system with its fractionally reserved banking system and free-floating paper currencies, gold is the only form of money that is safe from the depredations of central bankers. Moreover, it represents no-one’s liability and hence does not depend on anyone’s “promise to pay”. Its price trend in terms of paper money is best regarded as a barometer of confidence in the central bank administered monetary system.
That the US dollar has lost about 97% of its value against gold since the establishment of the Federal Reserve. This trend is certain to remain a one-way street over the long term, interrupted only by occasional fluctuations as confidence in central bank-issued paper (or digital) money waxes and wanes in cyclical fashion. This unfortunate state of affairs obviously doesn’t seem to worry the engineers of inflation. They only get concerned when their product loses purchasing power too quickly (i.e., when the debasement of fiat money happens at a pace that makes everybody take notice).
So what are the main factors driving gold demand that one should focus on? In no particular order, they are: the level of real interest rates (as determined by nominal rates minus market-based inflation expectations), the spread between short- and long-term interest rates (which serves as a proxy for the bias of monetary policy and/or changing inflation expectations), credit spreads, the rate of money supply growth, inflation expectations, the performance of “risk assets” and the exchange rate of the US dollar. An additional factor worth paying attention to (to our knowledge it was first mentioned by Steve Saville), is the relative performance of bank stocks. This is a sensible addition to the list, as bank stocks (and bank bonds) are among the assets reflecting confidence in systemic stability.
Real interest rates are relevant for the opportunity cost of holding gold. Given that gold has no yield, fiat currencies can only compete with it by offering an interest return. The steepness of the yield curve affects gold for two possible reasons: it can either steepen because loose monetary policy is pushing short-term rates down, or because rising inflation expectations are pushing long-term rates up – both are bullish for gold (conversely, the opposite is bearish for gold). Credit spreads are an indicator of economic confidence. If lower rated debt declines relative to higher rated debt, it indicates declining economic confidence, which is bullish for gold.
Credit spreads (Merrill Lynch high yield index spread) – a sharp decline in credit spreads between 2011 and 2014 proved to be bearish for gold.
Money supply growth informs us about the pace at which the fiat money supply grows relative to the gold supply. In the current monetary system, this is a permanently supportive factor for the gold price, only its intensity fluctuates over time. Inflation expectations, i.e., the expected decline in the purchasing power of fiat money are relevant for all prices, as market participants will tend to discount these expectations. The gold price tends to be especially sensitive to changes in these expectations. The performance of “risk assets” more generally (especially stocks) is also a reflection of economic confidence. Moreover, they compete with gold for the attention of investors. Lastly, the exchange rate of the dollar is important because all commodities, including gold, are mainly traded in dollars. If the dollar rises, gold will become more expensive in foreign currency terms (and vice versa if the dollar declines), and this will ceteris paribus pressure (or lift) gold prices in USD terms.
One needs to keep in mind that which one of these factors will exert the greatest influence on the gold price largely depends on contingent circumstances. For instance, there have been periods in recent years during which dollar strength didn’t put pressure on the gold price, due to concerns over the survival of the euro. Most of the time the factors listed above won’t be in perfect sync; at any given point in time, some will be bullish, while others will be bearish for gold. If a majority of them is bullishly aligned, the gold price is fairly certain to rise. Which factors should be given the greatest weight has to be judged on a case-by-case basis. Historical experience suggests that the level of real interest rates is especially important for the long term trend though.
Typically gold is a counter-cyclical asset that performs best in real terms when economic confidence and financial market liquidity evaporate and people desire to increase their savings and cash balances. Occasionally, there are also periods when pro-cyclical demand is in evidence and the prices of equities, commodities and gold all rise concurrently amid abundant liquidity. While strong gold price gains in nominal terms tend to occur during such periods, they are as a rule far less pronounced in real terms. In these phases silver can actually be expected to outperform gold, whereas gold will usually be the better performing asset in times of declining economic confidence.
To summarize: In order to explain and forecast movements in the gold price, one must focus on trends in monetary or investment demand. This requires observing and forecasting the trends of the relevant macro-economic data listed above.
As mentioned, gold ultimately serves as a barometer of the confidence of market participants in the government-run monetary system, and closely connected with this, also the fiscal situation of governments. The fiscal situation of the US government is especially important in this context. This is due to the fact that the dollar has become the world’s foremost trading and “reserve” currency. Partly, this is a remnant of the Bretton Woods system, during the operation of which the dollar was employed as a proxy for gold, however, it is also a result of the fact that the US remains the world’s primary economic and military power and is seen as the guarantor of the Western economic, monetary and political system. Anything that weakens it will obviously affect market confidence.
The above implies that many of the things people often focus on in explaining or forecasting gold price trends are at best of marginal importance. This includes data points such as whether central banks are net buyers or sellers of a few 100 tons per year, whether India imports 600 or 800 tons, and so forth. Similarly, whether gold moves from COMEX warehouses to warehouses in Shanghai is irrelevant to the gold price (this should actually be glaringly obvious by now, but apparently it still needs to be said now and then).
This is not to say that such data have no effect at all – they are just not important for the bigger picture and the long term trend and their effects are best of a psychological nature. In fact, such psychological effects can as a rule only be discerned if they are in line with an already prevailing price trend.
As an example for this, consider when it became known in early 2010 that India’s central bank had bought 200 tons of gold from the IMF. At the time, the gold price rallied by about $40 in a single day, seemingly in reaction to this news item. However, over the past three years, annual central bank purchases of gold have been far higher than in 2010. And yet, news of this increase in central bank buying have had absolutely no effect on the unfolding downtrend.
In assessing gold, one needs to eschew the usual commodity supply-analysis model. Gold is of course a commodity, but one with unique characteristics. No other commodity (except for silver, which is a hybrid between industrial and monetary commodity) has a remotely comparable stock-to-flow ratio. The sizable stock of above-ground gold is in fact a major reason why it is useful as a money commodity. Gold therefore needs to be analyzed as a currency – the only one that isn’t subject to the baggage that comes with the fiat currencies issued by governments.
This is so to speak a “Golden Oldie”, i.e., an expanded version of the first part of a two part article series we wrote for Mish’s Global Economic Trend Analysisin mid 2007, before this blog was established. We have discussed the above points on previous occasions, but decided to put it all together in a single post on the basis of the original article. In an edited version of the second part “Why Does Fiat Money Seemingly Work”, we will look at gold’s unarmed opponents in the world of money.