Supply and Demand. Boring until it’s not. March 11, 2018
In the treasury market, I’ve never really bought into the idea of ‘supply and demand’ as a driving factor. After all, we’ve previously been in periods of increased deficits (more supply) that don’t seem to impact rates much at all. I suppose one would argue that other factors were increasing demand at the same time, for example, perceptions of lower future growth. During episodes of QE, increased demand by central banks actually seemed to have more of an impact on stocks than bonds; stocks typically went higher and bonds edged lower. One might conclude that simplistic concepts of supply and demand have little to do with yields; influences are much more nuanced and intertwined. For example, deficits and supply often go up during economic downturns, but the same macro picture also leads to lower rates and demand for fixed income.
Currently we are in a period of rising rates, engineered by both Federal Reserve normalization policy, and increased supply driven by swelling deficits. There are many technical divergences that might suggest yields have pushed too high and are due for a substantial retracement. I’ll mention three.
First, the top chart indicates a break in the relationship of $/yen and the ten year yield, with the former moving lower and the latter pressing higher, a stark change since late 2017. Do we necessarily get ‘reversion to the mean’? This divergence might simply signal the idea that Japan may be moving towards the stimulus exit door leading to a stronger yen and less fixed income buying globally. If so, the reversion trade may easily transform into revulsion.
A second, though less glaring divergence concerns a relationship popularized by Jeffrey Gundlach of Doubleline. His arsenal of indicators includes the ratio of copper to gold versus the ten year yield. Makes sense, copper is related to economic activity and gold is a ‘store of value/flight to quality’, so as this ratio goes up it would generally correlate to ten year yields. The chart is below. The copper gold ratio peaked right at the end of 2017, but yields have continued to move higher.
Third, I consider the Commitment of Traders (COT) reports. I don’t personally weight this data heavily, because I think it’s a lot more valuable for commodities than financials, but I’ve cherry-picked a chart to include here anyway. The white line is the non-commercial (spec) net position, a record short in the Ultra Bond. The amber line is the continuous front ultra bond contract (WN1). Specs are hugely short. Does that necessarily lead to a rally? Dunno. **By the way, the chart is mis-labeled, Amber is the color of the futures price.**
Looking at these three indicators, one might determine that the rise in yields has gone too far. Reviewing the historical record, one could further support the above conclusion. For example, many commentators have compared the current environment to 1987. Mostly, they are talking about stocks, noting that the market rallied even in the face of rising rates, but then crashed. Here, I will just focus on the ten year yield. In early 1987, the ten year had bottomed just above 7%. The Fed was tightening, and the yield peaked in Sept 1987 just above 9.4%. Cumulative move was 240 bps, an increase in yield of about 35%. Yields then fell of course, in the aftermath of the crash. [Using data from St Louis Fed].
There was another notable rate increase in 1993/1994. In 1993 the Fed left the overnight rate at the then historic low of 3% for over a year. In 1994, an aggressive hiking schedule commenced. The ten year yield bottomed at 5.7% in Sept 1993 and topped just under 8% in Nov 1994. Cumulative move 230 bps, an increase of around 40%.
The tightening associated with 2004-2006 was much more drawn out. The ten year yield bottomed at 3.3% in June 2003 and topped at 5.1% in 2006, for a total of 180 bps, an increase of about 54%.
In the current cycle, the low yield associated with Brexit was just under 1.5%. The yield has just about doubled, closing on Friday at 2.89%. Shall we focus on the cumulative change of 140 bps and figure there’s another 100 bps to go in order to reach the 240 bp rise of 1987? Or do we say that a doubling of rates in an environment of huge debts is already punishingly overdone? By the way, this move is more or less a replay of 2012 to 2013, when the yield rose nearly 140 bps from August 2012, (1.68%) to the end of 2013, around 3%.
In the very short term, the Treasury auctions 3 and 10 year notes on Monday, followed by the 30 year bond on Tuesday. Given the resignation of Gary Cohn and an increase in protectionism, there is some concern that Asia might go on a buyers strike. If I were China, I would make a point by selling reserves this week. In the very short term, supply and demand takes on heightened importance. Also this week, CPI is released Tuesday with yoy Core expected 1.8%, same as last month. Retail Sales and PPI on Wednesday with PPI yoy Core expected +2.6%.
In the final analysis, yields are driven more by macroeconomic factors than treasury supply. We are in an environment where macro factors are lining up with supply considerations to produce a bearish outlook on both counts in spite of short term technicals. Friday’s strong nonfarm payroll increase of 313k didn’t spark a large sell-off because the market is already short, and was looking for confirmation of wage pressure with Average Hourly Earnings growth. YoY AHE was only +2.6%, a deceleration relative to last month’s 2.8% rate, so weaker shorts pared positions.
In conclusion, while there are several technical indicators that portend a short squeeze, I think inflationary concerns coupled with QT will dominate, and yields will continue to press higher. As an aside, while we know that demographics and debt where of a much different composition, I’ll bet the generation of 60 years ago, with an identical ten year yield of 2.9% in 1958, could scarcely have imagined the bond bear they were about to experience, culminating in a yield over 15% in 1981.
OTHER MARKET TIDBITS
It was a pretty big week for calendar spreads in Eurodollars. Near term spreads continued to press higher, but the back end of the curve compressed. As an example, EDZ8/EDZ9 spread rose 4 bps to 36.5, while EDZ9/EDZ0 fell 1.5 to 7.0. As mentioned, EDZ9 has become the elephant in the room, with open interest of 2.175 million. Since the start of 2018, the open interest in this contract has more than doubled! Heavy buying in EDU8/EDH9 spread from 24 to 25, settled 25.5. (This appears to be a roll of shorts to EDH9). Also heavy buying in EDM8/EDZ9 and EDZ8/EDZ9. The short in EDZ9 appears to be in strong hands. The peak one-yr spread is EDH8/H9 at 48.25. EDH8 expires in one week; EDM8/M9 is 47.0.
Over the previous 5 weeks, green midcurve straddles have not declined in premium at all. (Thanks for pointing that out BC). For example, On Feb 2, 2EM 9712.5^ settled 33.5 vs 9715.5 and on March 2 it settled 33.5 vs 9717.0. No time decay. On Friday, 2EM 9712.5^ settled 30.5 vs 9708. The long dated EDM20 straddle actually rose over the time period: 70.0 on Feb 2, 73.25 on March 2, and 76.25 on Friday!
_______________________________________________________
| 3/2/2018 | 3/9/2018 | chg | |
| UST 2Y | 223.4 | 226.2 | 2.8 |
| UST 5Y | 262.0 | 265.2 | 3.2 |
| UST 10Y | 285.3 | 289.2 | 3.9 |
| UST 30Y | 312.6 | 315.9 | 3.3 |
| GERM 2Y | -55.2 | -55.6 | -0.4 |
| GERM 10Y | 65.1 | 64.8 | -0.3 |
| JPN 30Y | 75.8 | 76.3 | 0.5 |
| EURO$ Z8/Z9 | 32.5 | 36.5 | 4.0 |
| EURO$ Z9/Z0 | 8.5 | 7.0 | -1.5 |
| EUR | 123.21 | 123.07 | -0.14 |
| CRUDE (1st cont) | 61.25 | 62.04 | 0.79 |
| SPX | 2691.25 | 2786.57 | 95.32 |
| VIX | 19.59 | 14.64 | -4.95 |
Note: One ED one-year spreads I switched to Dec contracts as they have the largest volume and open interest
