|
During the first nine years of this gold bull
market, gold prices moved with a nearly perfect inverse relationship to the
U.S. dollar. Indeed, in the early years gold was only really moving up
against the greenback; it was only after a few years that it began to
appreciate against all currencies. The game plan was simple; The dollar is
going down, so gold in USD terms is going up with some leverage factor.
Gold worked well as both a USD hedge and as a tool to speculate on a
USD decline ... but this is no longer the case.
 
Nothing lasts forever, and over the past two years or so this inverse
relationship has broken down significantly. The gold story is no longer
simply a USD devaluation play.
As the above chart shows, although there are times when the inverse
relationship remains intact, there are long periods where gold and the USD
move together.
The most significant reason for this is that the euro is a lot less
desirable than it was a few years ago. Since all currencies trade on a
relative basis, it doesn’t matter if the USD has poor fundamentals; if
the picture for the euro is worse relative to the USD, then the greenback
will make gains against the euro. During periods where the euro zone debt
crisis has been the focus of market attention, gold and the U.S. dollar have
been bought, since both are preferable to the euro and provided a safe haven.
The key point of this article is not to say that gold will not rise is
if the U.S. dollar falls; it is to point out that gold is no longer as
effective as a USD hedge. To show this, we present a scatter plot of the
closing prices for the USD index and gold over the last two years.
Although the trend line has a slightly negative slope, it is hardly
convincing, and the R-squared value of 0.0188 further diminishes the
creditworthiness of gold as a USD hedge. For those unfamiliar with this, the
R-Squared value is a statistic that indicates how good one variable is at
predicting the other. For our purposes, it is a measure of how good a decline
in the USD index is at predicting a rise in gold. If the R-Squared value is
1, then given the value of one variable, one can exactly predict the value of
the other. If R-squared is 0, that means that knowing the value of one
variable does not help you predict what the other variable will be.
So the higher the R-Squared value, the better one variable is at
predicting what the other may be -- and therefore the stronger the implied
relationship is between them. An R-squared value of 0.0188 is extremely poor,
and almost indicates no predictive abilities between the movements in the USD
and gold.
However, to get a fairer picture we should look at the relative
returns of the USD and gold.
This does give us a higher R-squared value at 0.0946, but it is still
very low and hardly convincing that gold has been an effective hedge against
declines in the USD over recent years. Repeating the above exercise for silver
and the USD index yields similar results, with R-squared values of 0.0844 and
0.1269. Although these are slightly higher than gold is, it is still nothing
to write home about -- let alone base a trading or investment strategy on.
To give you an idea of what a strong relationship between two assets
should look like, we have repeated the above exercise for gold against
silver, a relationship which is much stronger.
Although gold and silver obviously have a positive correlation versus
the negative correlation between gold and the USD, and silver and the USD, we
are not looking at whether the relationship is positive or negative; we are
only concerned with the strength of any such relationship.
Furthermore, when we calculate the correlation coefficients for gold
and the USD we get -0.137; for the USD and silver we get -0.290 -- both of
which imply a very weak negative correlation. Compare this with the very
strong positive correlation between gold and silver of 0.905.
Therefore, statically speaking, over the past two years gold has been
a very poor hedge against a declining USD. Of course this situation may
change, and just because the relationship has been weak over the last couple
of years doesn’t mean that there were not periods where the USD and
gold exhibited strong negative correlations.If the
USD index were to fall out of bed we would be expecting gold prices to rise;
however, over a broader horizon or when moves are more moderate or contained
within a range, don’t count on gold moving the opposite direction to
the greenback.
In our opinion, if you hold a view that the USD Index is going to
fall, then short the USD Index rather than taking a long position on gold.
You are running the risk that gold and USD could move together, and you are
not being compensated for that risk by any leverage factor in gold.
If you cannot trade futures, there are ETNs that allow you to gain
exposure to the USD Index, such as the PowerShares
DB US Dollar Bullish Fund (UUP)
and the PowerShares DB US Dollar Bearish Fund (UDN).
Options are traded on these funds as well, if you wished to use options to
play a move in the USD Index.
In conclusion, we hope to have shown that gold simply isn’t a
clean hedge against the USD anymore. When trading, one should aim to tailor
positions to match one's view and optimize the risk/reward dynamics. If you
think that the USD is going to fall, short it. If you think gold is going to
rise, then buy gold. Taking a long position on gold purely because you think
the USD Index is going to fall is not really statistically justifiable, due
to the weakness of the relationship between the two.
Disclosure: I am long GLD, SLV via a combination of option
holdings
Sam Kirtley
SKkoptionstrading
|
