Last week, I argued that the gold/oil ratio intimated a risk of a Leeb oil shock (an 80% year-on-year increase in WTI crude prices) by the end of the third quarter and that the risk was sufficient to be hedged against. This week, I am going to argue that the gold/oil ratio suggests a strong possibility that bonds are approaching a top.
Whether or not this is also the big top in the bond market that everyone has been trying to mark, I am not willing to venture here. Too many prophets have declared that, at long last, the time to sacrifice the epic bond bull has finally arrived, only for it to leap up and chase the doomsayers around the altar. Or, an angel comes out of the bushes at the last minute and asks, "What are you doing with that knife?" And, the precious metals market gets cut up instead.
What I am willing to say in this article is that yields are awfully low right now and that we have tools that suggest bonds may soon be under some pressure. If that is the case, investors may want to consider looking for points at which to shift into more interest rate-sensitive stocks and ETFs.
In other words, I merely intend to argue that the bond bull is, in the end, just a cow.
But, first let's try to get a grip on where we stand now.
Ten-year treasuries have historically tracked fairly well with the copper/gold ratio. And financials tend to track well with both of them.
There are certainly key moments when the relationship breaks down, but for the most part, it appears to hold.
To sharpen the correlation a little, I prefer to compare the ratio of ten-year yields to the dollar index with the copper/gold ratio. Where higher yields and copper suggest growth and risk-taking, the dollar and gold represent safety and risk-aversion.
This relationship has become so tight in the last decade that the copper/yield ratio mirrors the gold/dollar ratio in an approximation of Gibson's paradox - or at least until this year, when yields appeared to flag on a relative basis.
So far, this is fairly conventional, I think. If we zoom out, however, we find that the correlation between the copper/gold and ten-year yield doesn't appear to hold up very well. In fact, it seems to be perfectly backwards: over the longer term, the gold/copper ratio seems to be correlated with yields fairly strongly instead.
And the solution to this problem of how gold/copper and copper/gold could both have fairly strong correlations with yields is, I believe, to shift the gold/copper ratio forward sixteen months so that it "forecasts" (however roughly) yields and, somewhat perversely, the copper/gold ratio. I think this is an example of what I would call "intermarket reflexivity," the tendency for a trend to generate its antithesis.
Although a proper economist would probably be better fit to consider the question, this raises the issue of why the secular rise of the copper/gold ratio (i.e., the fall of the gold/copper ratio), representing economic dynamism, should coincide with a secular fall in interest rates, representing safety. My first instinct is to blame the perversities of our monetary system and the recycling of dollars as described by Richard Duncan in The Dollar Crisis, but it would require more careful consideration.
In any case, generally speaking, much of what I have said regarding the relationship between the gold/copper ratios and yields holds true for the relationship between the gold/oil ratios and yields.
(click to enlarge)One or the other of these two ratios, or some third one (perhaps silver/oil), may in the end prove to be more fit to forecast yields, but I am particularly interested in the oil/gold ratio and I have more data and a better feel for it, so I am going to stick with it here.
Here is the gold/oil ratio shifted forward sixteen months against ten-year and three-month yields and their spreads.
The magnitudes of the moves do not appear to fit especially well, but a relationship certainly seems to exist. If you look closely, the periods where the correlation between the forward gold/oil ratio and the ten-year yield breaks down tend to be followed fairly quickly by a significant flattening of the yield curve, generally to a point where the spread falls below 1.0, a level that I argued was significant for predicting oil shocks.
In fact, although I don't want to overdo this, historically, we can model this to some degree.
From the point that the correlation enters negative territory, the yield spread tends to fall under 1.0 within a matter of months.
Beginning around 2004, around the time of Greenspan's "conundrum," the correlation gave way and then went strongly negative. After threatening to go positive again in the wake of the financial crisis, it has again gone strongly negative.
Even so, the forward-looking relationship between gold/oil and yields has managed to keep much of its charm recently. I prepared the following two charts almost a year ago when I first began considering these particular market dynamics.
The 14.28 level (0.07 in oil/gold terms) appears to be significant again.
That has brought us to today, where we appear to be approaching point 'G' sometime around August.
And, then looking out over a possible future of rising yields until the end of 2012, indicated by the spike in gold/oil that occurred in the summer of last year.
The thrust of my argument is similar to my previous argument about oil. First, the history of the gold/oil ratio suggests that yields are too low. Second, despite the strong negative correlations of the last decade, the gold/oil ratio still seems to be signaling trends in yields sixteen months in advance. Third, the gold/oil ratio seems to be signaling a trough sometime around August, followed by a spike around the end of 2012.
These points are sufficient, I think, to start giving treasuries a wider berth and to seek out exposure to interest rate-sensitive vehicles such as XLF or financial stocks that historically track well with interest rates or the copper/gold ratio. One could also consider going long the Nikkei (NKY) and/or shorting the yen (FXY).
In my next piece, I would like to look at bonds with reference to another historical pattern to consider how close we might be to the end of the bull market in treasuries.
Disclosure: I am long BAC.
Additional disclosure: I am long September WTI.