Friday the 13th seemed as good a day as any to clean up shop a bit and do something as boring as recapping our arguments, as well as doing something as foolish as trying to weave them into a single narrative. I can't remember any such attempt on my part having panned out in the end, and in general I try to get each argument to stand on its own merits, especially when it comes to making decisions about where to put my money. On the other hand, forcing ourselves to synthesize our various positions and outlooks has the benefit of exposing our ignorance, although hopefully not only to others.
So, here is where we stand now:
- Based on movements in the gold/oil ratio, we should be prepared for a possible "Leeb shock" (an 80% yoy increase in crude prices) sometime this summer, most likely August. The yield spread, however, has not confirmed that, so it is a tentative call, and last month I recommended that crude should be bought when it challenges the 2011 lows. So far, that call has not resulted in ruin.
- Based on movements in the gold/oil ratio, I argued that bonds are likely approaching at least a temporary top this summer, and that they would then slide the rest of the year.
- Based on #2, I recommended getting into cyclical stocks, but partly due to a reader's challenge, I sharply reversed a bullish stock market outlook, primarily because the gold/copper ratio warned of a significant rise in real interest rates over the latter half of 2012, especially in the fourth quarter.
- Based on the five-year cyclicality of interest rates, we should expect interest rates to bottom in 2012 or 2013, but that bottom should be followed by a strong snap-back.
Okay, so let's start from the bottom, and work our way back up.
click to enlarge images
This chart allows us to try to tie some pieces together, although a deeper accounting for the phenomena still eludes me.
The spikes in the gold/copper ratio tend to coincide with the (5-year) cyclical lows of interest rates, in what almost appears as an attempt by the classical version of Gibson's Paradox to reassert itself.
In the table below, I try to chronologically synthesize Reinhart and Rogoff's major post-Depression financial crises with Leeb shocks and with gold/copper and cyclical patterns.
Reinhart & Rogoff, Leeb, and Gibson's Revenge?
|Year||Crisis||Cyclical low||Gold/copper "shock"|| |
|1981-2||global (Latin America)||1981-2||1980||1980-1|
|1987-8||Norway; global (Africa)||1987||1986||1987|
|1991-2||Finland, Sweden, Japan; global||1993*||No||1990|
|1997-9||global (Asian Contagion)||1998||No||No|
|2007-?||global (Great Recession)||2008||2008-9||2008^|
|2012-2013?||Something new or something old?||2012?||2011|| |
* In the case of 1993, I am not sure which crisis to attribute the cyclical low to, so I list it with both.
^ These Leeb shocks occurred after the publication of Leeb's book and are based on my own calculations. Also, the clustering of financial crises makes it difficult to identify which might be associated with which events.
Based on this table and referring back to my arguments above, it seems at least not wholly implausible that a Leeb shock could occur in the middle of a financial crisis. In fact, that may be a major contributing factor to these shocks in the first place. This would also suggest that the jumps in gold/industrial commodity ratios (e.g., gold/copper and gold/oil) that appear to 'predict' Leeb shocks and interest rate movements contain valuable information. But, they require a lot more theory before anything conclusive can be said, and there are countless questions to be answered: Why are there sixteen-month 'lags'? Why is there a five-year cyclicality? Why should Gibson's Paradox take such a strange form? Why should oil shocks occur in the middle of banking crises?).
Moreover, as I mentioned above, if the market is consciously or subconsciously pricing in these events a good year in advance, why can't it price in, consciously or not, the results? In other words, as I pointed out above, Leeb shocks--which at least appear to force stocks to capitulate--are foreshadowed by things like the gold/oil ratio and the yield curve, so why should equities run up to a great crescendo and then collapse, such as in 1987 or 1999 or 2008 amid shock and dismay? Unless, of course, the patterns I am pointing to are imaginary (always possible) or there is a deeper structural dynamic generating these sequences (unlikely but probable, if you will).
Let's try to isolate one more piece of the puzzle before trying to relate it all back to the future. The gold/oil ratio tends to indicate interest rate movements sixteen months in advance (not unlike the gold/copper relationship with real interest rates) and Leeb shocks twelve. Yield spreads below 1 tend to indicate Leeb shocks sixteen months in advance.
What does this say?
Well, for one, that the yield spread should give us a preview of nominal interest rates sixteen months in advance.
There seems to be something of a lead here, but it is not easy to use it for timing the market or to gauge magnitudes. However, if we look at it just with reference to the five-year cycles, the yield spread appears to peak very distinctly, although roughly (okay, very roughly), sixteen months prior to a bottom in yields. Although my chart only goes up to 2011 with the yield spread, it has crept down (up on this chart) to 1.5, which gives the impression that we are very close to a top in the bond market, although it doesn't tell us which top, short- (six to twelve months), medium- (five years), or long-term (a decades).
But, there may be information about the Big Top's arrival when it comes: Looking out over the last 30 years, most cyclical lows are followed by a snap-back that sets a high that is never attained again (usually in a year ending with '0' or '5'). If this pattern should continue, we can expect a snap-back after the cyclical lows we are approaching now. If interest rates should move above the 2010 highs (about 4%), that might provide an important clue that the game has changed. Otherwise, we might expect the next cyclical high to constitute the new line of resistance.
Moving back to the present, the yield spread gives the impression that the cyclical low should be upon us already. But, one cause for hesitation is that the spike in gold/oil of last year occurred at an usual time in the five-year cycle. Gold/oil, like yields, tends to function on a similar kind of cyclicality, as you may be able to notice from the following chart.
Most significant spikes in the gold/oil ratio occur sixteen months prior to the cyclical peak in yields, so it is unusual to have a spike in a year ending with a '1' or '6', although this did happen in 1986 in conjunction with an unusually early cyclical low in yields. From a cyclical point of view, 2011 should have been seen setting significant lows in the gold/oil ratio (rather than highs). And, in fact, it did, but they were among the least impressive lows on record, only barely reaching below the key 14.28 level and the previous years lows--soon after which, the spike I am making so much of occurred.
So, like everything else in the New Normal, nothing is going to be easy, but much of this suggests that the short-term bottom in yields we can expect soon (my sense is some time around the end of this month) might also turn out to be an early cyclical low. The yield curve suggests it, and so do the gold/commodity ratios. It is hardly a certainty, but the odds seem to be rising that this is a significant top approaching.
In sum, this suggests that one should not only be unloading Treasuries but preparing to short them (IEF). Identifying the five-year cyclical low is important, because of the rapid rise in interest rates that typically follow those lows (the 'snap-back', as I've been calling it). Acquiring exposure to that possibility seems increasingly reasonable.
The stock market is just as likely to be flat or lose ground. To give an idea of what the future holds, I am placing a chart of the copper/gold ratio from sixteen months ago over a current chart of the S&P 500 (SPY). Although the magnitudes of the copper/gold ratio shouldn't be taken into account in this instance, it gives an idea of how momentum in the copper/gold ratio sixteen months in advance leads stocks.
And, as I argue, how equities will roughly look over the next sixteen months, counting from March 2011 (again, the magnitudes I do not believe are significant):
The fall in this ratio during the first three quarters of 2011 suggests weakness in stocks from June until the early months of 2013. History also suggests that this weakness may be nothing more than a flat market rather than anything catastrophic.
Betting on stronger volatility (VXX) may provide one of the safest ports in this gathering storm. Likewise, shorting the Aussie dollar (FXA) against the yen (FXY) or the Swiss franc (FXF) might offer protection against a market that struggles to find direction over the course of the latter half of 2012.
Finally, one last discussion to consider how a scenario of both rising nominal rates and real rates could sit beside a bet on rising crude oil in what seems like a deflationary environment.
The following chart appears to be a sibling of Gibson's Paradox, as I've tried to reformulate it.
Rather than the gold/copper ratio's 'prediction' of real interest rates sixteen months in advance, this ratio seems to forecast changes in inflation twelve months in advance.
Here is the silver/oil ratio looking back over the last three years.
This suggests that inflation should be rising into August 2012 and then collapsing. This seems to point again to a potential rise in year-on-year crude prices, and interest rates often track these year-on-year rises, but they often can take longer to retreat when those year-on-year prices fall.
Above are the last two Leeb shocks. In the first case (2008), rates didn't begin to really fall until months after yoy oil fell, and in the second (2010), they fell largely in sympathy, but even then they peaked after yoy oil did.
And below is where this relationship stands today.
I would argue that any jump in August prices (particularly oil) will have a disproportionately strong effect on CPI rates, because of the collapse in commodity prices that time last year (August 2011), and that this could lead to a jump in interest rates that would fail to fall in sync with rapidly falling inflation rates, thus creating a rise in real rates hard on the heels of the rise in nominal rates.
In this environment, it is not difficult to imagine market pressure causing volatility and monetary tightness, exacerbating the ongoing financial crisis and perhaps pushing yields down again into 2013. We could hardly be surprised at such an outcome, but at the moment, 2012 looks like a better bet for the top in bonds.
We'll see. In these kinds of narratives, it usually only takes one thing to go wrong for the whole thing to unravel. The biggest intermediate marker will likely be where things stand at the end of August.
Disclosure: I am long September WTI and short AUDCHF, AUDJPY, and September S&P 500 futures.