Introduction
Most analyses of the gold market consider the annual
change in the amount of gold produced by the mining industry to be an
important determinant of the gold price, with bulls regularly supporting
their case by citing the mining industry's inability to ramp up production
and bears sometimes claiming that increasing mine production will eventually
weigh the gold price down. Our contention, however, is that the annual supply
of newly-mined gold is so small relative to the existing aboveground supply
that changes in mine production should be ignored when assessing gold's
prospects.
We warn you that the following discussion is quite
lengthy (by our standards), but we wanted to cover most aspects of this
important topic -- important, that is, for anyone who genuinely wants to
understand the gold market -- in one hit.
The aboveground gold supply (and why it makes sense
to analyse gold as if it were a currency)
The gold that gets produced each year doesn't get
consumed; rather, it gets added to the existing aboveground stock. This means
that the entire aboveground stock represents potential gold supply. In this
respect, gold is more like a currency than a commodity.
Now, we realise that not
all the gold that was ever mined is currently in a readily saleable form, but
a substantial chunk of it is. To be more specific, it has been roughly
estimated that around 150,000 tonnes of gold have
been mined throughout history, about two-thirds of which has been mined since
1945. Of this 150,000 tonnes, we estimate that
around 108,000 tonnes* are held for
monetary/investment purposes. Monetary/investment (MI) gold includes official
gold holdings (the gold kept by central banks and the IMF), privately-held
bars and coins, the gold held by ETFs and closed-end funds, and monetary jewellery (24-carat jewellery
that is held solely as an investment or a store of value). This 108,000-tonne
figure is, in our opinion, a rough but reasonable estimate of the MINIMUM
amount of gold in readily saleable form, and clearly dwarfs the 2,400 tonnes of gold produced by the mining industry over the
past year.
Analysing
the gold market as if new mine supply dominated the supply side of the
equation would be like analysing the dollar market
as if the only dollars that really mattered were the new dollars that came
into existence over the past year. The fact is that a new dollar created
today will become an indistinguishable part of a total supply that includes
every dollar created, but not yet extinguished/destroyed, up until today, and
it is this total supply, combined with the associated demand for this total
supply, that determines the dollar's price. Moreover, it could be argued that
the gold mined each year is less important to gold's supply-demand balance
than are the dollars created each year, because new dollar supply typically
constitutes a much greater percentage of the existing stock than does
newly-mined gold supply. Specifically, whereas the total aboveground supply
of saleable gold increases by about 2% every year, the total supply of
dollars sometimes increases at a double-digit rate and rarely increases by
less than 5%.
Putting things into perspective
As discussed above, the total supply of gold in
readily saleable form is probably at least 108,000 tonnes,
which stands in stark contrast to the current annual mine supply of around
2,400 tonnes. This suggests that changes in
investment demand (changes in the demand for the total aboveground stock) are
more than 40-times more important to gold's price trend than changes in mine
supply. To put it another way, a 0.25% (one quarter of one percent) change in
investment demand is more important than a 10% change in mine supply.
Data provided by the London Bullion Market
Association (LBMA) lend additional support to the argument that changes in
annual mine supply are tiny in relation to the overall market. The LBMA
reports that an average of around 20M ounces (650 tonnes)
of physical gold changes hands on the London gold market every day. This
means that the equivalent of more than one year's mine supply changes hands
in the space of only four average trading days on the London market. And
London is not the world's only market for gold bullion.
If changes in mine supply are irrelevant, then what
is relevant?
Over periods of 2 years or less, gold's price trend
is usually determined by the combination of the US dollar's exchange value,
credit spreads, nominal interest rates, inflation expectations, and the yield
curve. However, the long-term is the focal point as far as this discussion is
concerned.
Our view is that long-term trends in the gold price
are driven by changes in the overall level of confidence in the monetary
system and the economy, as best indicated by the long-term trend of the broad
stock market (note: there is no reason why the stock market should be an
indicator of monetary confidence, except that since the 1930s governments
have generally implemented policies that debase the currency and create
uncertainty whenever the economy weakens). We point out, for example, that
the equity bear market of 1966-1982 coincided with a bull market in gold and
gold-related investments (gold stocks); that the equity bull market of
1982-2000 coincided with a bear market in gold and gold-related investments;
and that the equity bear market that began in 2000 has, to date, coincided
with a bull market in gold and gold-related investments. Correlation doesn't
imply causation, but it makes sense that the world's favourite
repository of savings over the ultra-long-term would tend to trend inversely
to the more speculative investments.
An implication of the above is that almost
regardless of anything else, gold's current bull market will continue until
the current equity bear market reaches its conclusion, which, based on
historical evidence, won't happen until after the average P/E ratio has
dropped to single digits and the average dividend yield has moved above 5%.
Wait a minute: most other commodities also rallied
during the 1970s, trended lower during 1980-2001, and trended upward during
much of the 2000s. How, then, does gold's strong tendency to move counter to
long-term trends in the broad stock market differentiate it from any other
commodity? The answer is contained in the following monthly chart of the
gold/CRB ratio.
To see the true picture it is often helpful to
remove the US$ (or any other fiat currency) from the equation by monitoring
the performance of things in terms of gold (the Dow/gold ratio being a popular
example) or by monitoring gold's performance in terms of other things.
Specifically, to find out what gold REALLY did during any period we can
review how it performed in terms of commodities in general (as represented by
the CRB Index). The following chart makes the point that the gold/CRB ratio
has moved counter to long-term trends in the broad stock market, meaning that
gold has out-performed most other commodities during long-term equity bear
markets and under-performed them during long-term equity bull markets.
Gold's performance relative to other commodities
during the 1930s is also worth mentioning. Whereas most commodities and
commodity-related equities did poorly during the massive equity bear market
of 1929-1938, gold and gold stocks fared extremely well.
Other factors that affect, or are widely believed to
affect, gold's price trend
Although our main purpose today is to deal with the
ramifications (or lack thereof) for the gold market of changes in gold
production, we will take this opportunity to also quickly deal with the
effects of central bank gold sales and changes in jewellery
demand. In particular, we want to quickly explain why the latter are
irrelevant and why the former have some significance, but nowhere near as
much as commonly believed.
The effects of central bank (CB) gold sales
In point form, here is a
summary of the central-banking community's influence on the gold market. Note
that we place the gold held by the IMF and the gold held by government
treasury departments (in the US it is the Treasury, not the Fed, that owns
the gold reserve) under the 'central-banking umbrella'.
1. CBs hold about 30% of the MI gold stock, so they
have the ability to exert influence over gold's short- and intermediate-term
price trends. However, their actual sales over the past 20 years have been
too small to matter. Specifically, the World Gold Council (WGC) reports that
CBs reduced their collective gold reserve at the rate of around
250-tonnes/year during the 1990s and 400-tonnes/year during the first 8 years
of the current decade. (By the way, this means that CBs reduced their gold
holdings at a faster rate during the current bull market than during the
final decade of the preceding bear market, which is consistent with our view
that their sales have not been a significant influence on the price trend.)
2. News relating to CB gold sales often has a
short-term effect, but does not appear to have altered intermediate-term
trends and cannot, in our opinion, alter long-term trends.
3. CB gold sales represent reduced demand by some
holders of the MI stock, but these sales could actually cause an increase in
overall MI demand. This is because confidence in a fiat currency could fall
if the gold reserves 'backing' that currency were reduced, especially during
a period when confidence was already in a fragile state for other reasons.
Something along these lines occurred during the second half of the 1970s,
when gold sales by the Fed (on behalf of the US Treasury) and the IMF were
quickly followed by increases in the gold price.
4. CB monetary machinations (manipulations of
interest rates, money supply, and pretty much everything else to do with
money and credit) are probably of far greater importance to gold's long-term
price trend than the gold sales/purchases they happen to make from time to
time.
The effects of changes in jewellery
demand
Many gold market analysts attribute great importance
to changes in jewellery demand. For example, we
occasionally read analyses where it is stated that jewellery
demand is something like 60% of total gold demand, but what they really mean
is that jewellery demand is 60% of the flow of new
gold (primarily mine supply, but also including scrap supply). In other
words, such analyses completely ignore the huge aboveground supply of gold
and the associated demand for the aboveground supply.
The fact is that changes in annual jewellery demand are even less significant than changes
in mine supply, and should therefore be ignored when assessing gold's
prospects.
Anticipating some objections
1. An argument we've come across is that at any
given time there may be few, or perhaps even no, sellers of the existing
aboveground gold supply, resulting in the market being dominated by the
regular selling of gold miners (gold miners continually sell at whatever the
market price happens to be at the time the gold is removed from the ground).
Well, we know from the LBMA statistics that the
selling of gold miners is small compared to the amount of physical gold that
routinely trades via only the London market. However, even if all current
owners of gold decided to 'sit' on their investment in anticipation of higher
prices the selling by the gold mining industry wouldn't have an outsized, or
even a meaningful, effect on gold's price trend. The reason can best be
explained by considering the hypothetical example of Bill and Fred, two
shareholders of XYZ Corporation. Bill owns 70% of the outstanding XYZ shares
and Fred owns 1% of the outstanding shares. In our example, Fred decides to
sell his 1% stake immediately while Bill plans to wait for a much higher
price before parting with any of his shares. Even though Fred is the one
selling in the present, Bill's decision to hold his shares off the market is
vastly more important, as far as the market's price discovery is concerned,
than Fred's decision to immediately sell. The reason, of course, is that Bill
controls 70-times more shares than Fred. Robert Blumen's
article at http://mises.org/story/3593
goes more deeply into this concept.
2. It could be argued that industrial metals such as
copper also have huge aboveground supplies if we account for the metal
'stored' in buildings and other structures. The difference is that this metal
is not in saleable form. For example, almost regardless of the copper price
it will never be economically feasible to tear down functional office and
apartment buildings for the sole purpose of extracting the contained copper,
meaning that the copper wiring and plumbing built into these structures
should not be counted as part of the aboveground supply.
Gold is the only commodity where the
"stocks-to-flow" or "stocks-to-use" ratio is so large
that the "flow" component can be ignored. This point was touched on
in an article posted by Steve Matthews, a hedge fund commodity strategist, at
http://www.lbma.org.uk/docs/alchemist/alch32_commodity.pdf. Mr. Matthews concludes: "It's fair to say that nothing else
even comes close to gold's 7,019 days [of remaining supply]. This leads us to
a proposition that I'm sure some of you have thought about before: the right
way to trade gold is as a foreign currency, not as a commodity. You would
need someone else to give you a trading recommendation; I abdicate my duty to
declare myself bullish or bearish flat price in the face of what I consider
overwhelming evidence that gold resides outside my supply/demand analytical
framework."
By the way, we think the 7,019 days of remaining
supply (the number of days of supply in aboveground storage, assuming that
there is no further production and that non-investment demand proceeds at the
current rate) mentioned by Mr. Matthews greatly understates the true
situation. Our assessment is that the days of remaining supply in the gold
market is at least 13,000 (about 35 years). In comparison, the days of
remaining supply in the copper market is typically in the 30-90 range.
3. There has been a loose inverse relationship
between gold production and gold's price trend over the past 40 years. To be
more specific, the following chart of world annual gold production shows a
downturn during the first half of the 1970s (a bullish period for gold), a
steady upward trend from the early-1980s through to around 2000 (a generally
bearish period for gold), and then a tapering off during the current bull
market. This has been interpreted as evidence that changes in mine supply do,
contrary to our analysis, have a material effect on the gold market. However,
such interpretations reveal the danger of blindly assuming that correlation
implies causation.
Given the size of the overall gold market it is not
reasonable to conclude that the production trends illustrated by the
following chart were important influences on gold's price trend over the
period in question. So how, then, do we explain the apparent negative
correlation between price and production that has arisen over the past 4
decades?
The most logical explanation is that there is, in
fact, a cause-effect relationship between mine supply and price, but that
price is the cause and mine supply is the effect (a higher price leads to
higher production, etc.). Due to the extremely long delay between cause and
effect, what we get is the appearance that falling production pushes the
price up and that rising production pushes it down; but what is actually
happening is that today's production levels are a response to price changes
that occurred at least a decade earlier.
The time delay is so long because it will generally
take at least a few years of higher gold-mining profit margins to stimulate
an increase in exploration activity, that will, after several more years,
lead to an increase in production; and it will generally take many years of
low gold-mining profit margins to prompt the gold mining industry to reduce
its production and exploration activity.
With regard to the past few decades, our assessment
is that rising costs combined with a fixed gold price during the 1960s
probably brought about the decline in gold production during the first half
of the 1970s, while the large increase in gold-mining profit margins during
the 1970s prompted increased exploration activity that eventually led to more
mines being constructed and higher production during the 1980s.
With gold-mining profit margins generally remaining
robust and with many people remaining convinced that gold's next bull market
was just around the corner, the industry continued its expansion throughout
the 1980s. It wasn't until the second half of the 1990s that the downward
drift in gold-mining profitability really began to take its toll, causing
exploration activity to all but cease. The result has been an essentially
flat production profile from the late 1990s through to the present day, but
note that the past 15 months' surge in the gold/CRB ratio (a proxy for
gold-mining profit margins) should lead to more aggressive exploration, and,
eventually, to higher gold production.
(Chart data from USGS for all years prior to 2006 and from the World Gold Council for 2006
onward)
Conclusion
Gold market analyses -- such as those put together by
Gold Fields Mineral Services (GFMS) -- that treat new mine production as if
it represented a substantial chunk of the total gold supply, and/or that
place great importance on factors such as jewellery
demand and scrap supply, are of no use to anyone whose goal is to understand
what drives the gold price. In fact, such analyses are worse than useless
because they create a false impression. The reality is that the contribution
to total gold supply made by newly-mined gold is so small that changes in
mine production should be considered irrelevant when assessing gold's upside
potential relative to its downside risk.
*The 108,000-tonne figure is derived from the
analysis presented by James Turk in his 1993 booklet "Do Central Banks
Control the Gold Market". In this booklet Mr. Turk estimated that the
amount of monetary/investment (MI) gold was 72.7% of the total aboveground
gold supply at the time (1993). To arrive at the 108,000-tonne estimate for
the current stock of MI gold, we have assumed that the ratio of MI gold to
total aboveground gold is the same now as it was in 1993.
Steve Saville
www.speculative-investor.com
|