Gold and money still remain the great intellectual debate in the economics circles with polarized sides that, on one end (hard money), argue that absence of gold as the monetary anchor is responsible for the loss of dollar’s value over time, while others (fiat money) argue that anchoring the dollar to the gold would unnecessarily constrain the monetary policy in times of crisis, particularly during episodes of debt deflation, so that the economy would unduly suffer because of inability to deliver sharp monetary easing.
We see these two intellectual schools unfolding, and not just along political lines, but also in the gold market, monetary policy and the demand for money.
Gold price has been rising rapidly, particularly since circa 2003 and the broad spectrum of the hard money folks would say that the reason for such spectacular rise in gold price is because the Fed is pumping in an unprecedented amount of fiat currency.
The fiat money folks often lack an explanation as to why gold is spiking so much but, to their credit, that lack of explanation does not mean that there isn’t one. Instead, fiat folks say that money printing did escalate since circa 2008 but that is as a response to a collapse in the velocity of money: in October of 2007, the velocity index stood at 10.367 and as of January 2012 the index is at 6.958, a 32.9% decrease in the velocity of money.
By comparison, the M1 money supply measure in October of 2007 was $1376.3 billion while in January 2012 the M1 stood at $2216.3 billion, a 37.9% increase – roughly equal to the rate of fall of the money velocity.
A 2010 World Gold Council (WGC) study, however, shows that spikes in M1, among other things, are positively correlated with the price of gold. In fact, WGC says that M1 is directly responsible for gold price effects.
“The analysis suggests that a 1% change in money supply in the United States six month ago produces, on average, a 0.9% increase in the price of gold today, assuming the money supply in the other regions does not change,” writes WGC.
A powerful conclusion like this typically conjures up a perception that a 1% spike in M1 spikes gold 0.9% because M1 rarely goes lower. In fact, out of 449 datapoints on M1 since January 1975, only 97 of them show M1 contraction, and this 21.6% contractionary (78.4% expansionary) occurrence is mostly concentrated in times when the Fed is looking for ways to constrain the red-hot economy.
So, if M1 is expanding in 78.4% of instances, then what causes the gold price to go negative or flat such as it did in the vast 20 year period from circa 1985-2003?
Irrespective of the answer, the very presence of this question casts some doubt at the findings that M1 has such a large, almost 1-for-1 influence on the price of gold.
Another variable to look at is the effect of the velocity of money on gold. It is reasonable to hypothesize, along the hard money thought, that an increase in the velocity of money would spike the gold price but the data does not always correspond to this hypothesis.
For example, from October 1993 to October 2000, the velocity of money was on a sharp upswing going from 6.071 to 9.267 yet the price of gold went from $355.5 to $273.7.
WGC study, however, assigns predictive power to the gold price.
“A simple empirical regression model tells us that a 10% increase in the price of gold tends to increase the velocity of money in the US by about 0.4% in 12 months time,” finds WGC.
This finding suggests that the hypothesis that money velocity impacts gold is wrong and instead needs to be reversed – that gold price change impacts monetary velocity.
This, of course, is again contrary to the 1993-2000 data on these two variables because the gold price change during that period tended to be negative while M1 was rising.
More importantly, the conclusion that gold price change has something to do with the change in M1 rate of change is problematic on the fundamental level: it is like arguing that the change in the speedometer causes your foot to press harder on the pedal. Mathematically, the statement is totally wrong because it says that the first derivative determines the behavior of the second one.
At this point, of course, we are entering the great intellectual realm of hard vs. fiat money with hard money folks, like the Austrian economists, who would tend to argue that inflationary expectations is what drives the the price of gold. Again, just because fiat money people are mute on this point it does not mean that there is no explanation.
Having said all this, if we take the data that WGC is basing its conclusions – gold price and M1 measures from January 1975 to now – and break up the data into various periods, we can find different levels of correlation.
For the overall period, for example, the correlation coefficient is 0.73 meaning that M1 and gold price move in the same direction 73% of the time. However, during certain periods the correlation is different.
For example, for the 1975-1986 period, the correlation coefficient is 0.48 meaning that 48% of the time the gold price has moved in the same direction as the M1. However, between 1987-1998, the correlation coefficient is -0.48 meaning that 48% of the time gold went into the opposite direction from M1.
Since breaking up the data different ways yields different correlation coefficients, the results are very intriguing if the data is broken into 2 periods, from January 1975 to March 2003, and April 2003 to present.
For January 1975 to March 2003, correlation coefficient between M1 and gold was a small 0.16 while from April 2003 onwards the correlation jumps to a remarkable 0.95.
Now, April 2003 is not an arbitrarily chosen point but a rather deliberate one because it is the date that the first gold ETF began its full month’s operation.
On March 28, 2003, Gold Bullion Securities listed the first ETF on the Australian Stock Exchange, an event that was mimicked well in other large markets several times over.
Since then, gold price has been on a tear taking a pause only during the darkest days of the recent Great Recession. Yet since, gold has been disjointed completely from the velocity of money and has been following the growth of M1 or the rise in availability of cash which now has an unprecedented access to “imaginary” gold via the ETFs.
During this period, the velocity of money has been collapsing so the model where gold price has some predictive power over money velocity appears as a gobbly-gook.
More significantly, the data seems to suggest that synthetic financial products like the ETFs that offer non-physical gold (i.e. imaginary) have finally linked the M1 to gold where such link exhibited only a mild correlation prior to the their appearance.
The cost of such linkage have been mounting since 2003, and not just because of gold, but because other commodities, particularly oil, are being traded under imaginary ownership terms and amount to a massive wealth transfer from consumers to financial syndicates that are peddling such bets.