Is this 2000 again? Yield curve inversion 2019!
Thursday’s trading session marked an event of significant importance: for the first time in a decade, a vast portion of the treasury bond yield curve is now fully inverted. We witnessed a slight inversion between five and two year rates in December, but in Thursday’s trading the bottom fell out of the entire intermediate portion of the curve.
The 1 year now prices at a premium to the 2, 3, 5, and 7 year treasury notes, and is very close to the yield of the 10 year maturity. The market now appears to be signaling it anticipates the Fed to begin cutting rates sometime late in 2019 or early 2020:
Needless to say, this is somewhat troubling. In the coming weeks, you will see more and more speculation about whether or not the US economy is headed towards recession.
I cannot say whether or not a recession is imminent (as no one can); however, the inverted yield curve now suggests this is a legitimate possibility. For a primer on yield curve inversions and what all of this means, read the yield curve for dummies.
The 5 Year / 3 Month Inversion
The ‘gold standard’ for inversions with no false positive recession readings in its history is the 10 Year / 3 month spread. As of this writing sits it a paltry 14.5 bps after being at 89 bps as recently as early November. Quite a turn around in the sentiment from the euphoria of the fall.
The 5-year 3-month (I’ll just call it the 5/3) nevertheless does have some predictive ability of its own, which shouldn’t be surprising as the 5 and 10 year notes are heavily correlated. The 5/3 spread against eventual recession dates is plotted below:
Since 1982, there have now been 5 5/3 inversions:
- June 89, recession followed in July 1990 (13 months)
- September 98, no recession.
- August 2000, recession followed in March 2001 (7 months)
- August 2006, recession followed in December 2007 (16 months)
- January 2019, ???????
PMI craps the bed, Initial Unemployment Claims Continue To Increase
Today’s selloff was no irrational tantrum but underpinned by continued deterioration in global economic data. Joining the fray was a continued slide in JP Morgan’s Global Manufacturing Purchasing Managers Index (PMI), which clocked in at the lowest level since September 2016.
As initially cited in December, the initial unemployment claims trend is rather worrisome for the US as well. More people than expected signed up for unemployment benefits, potentially signaling the first cracks of weakness in the labor market.
Is This 2000?
I’ve dedicated a lot of time analyzing the 2001 recession and looking for clues in the data as it would have been available and I’ve concluded you simply could not have seen a recession coming. It remains a huge part of my underlying economic philosophy, in which the dynamism and feedback is too complex, and the future is itself constantly changing.
But 2001 is interesting because the data was all fine and dandy right until it just wasn’t fine anymore. There was no Bear Stearns or Lehman Brothers moment where everything came crashing down, things just gradually kind of went into the crapper with a whimper. Next thing you know, boom, a recession is at the doorstep like a couple of Jehovah’s Witnesses at dinner time.
The current environment reminds me quite a bit of 2000-01 in terms of the yield spreads and unemployment data.
4 months separated the minimum of unemployment claims hit in April 2000 and the subsequent yield curve inversion in August 2000. More recently, unemployment claims bottomed in September 2018, and an inversion has again followed 4 months later.
When overlaid upon each other, the series have an uncanny resemblance. In fact we are exactly at an identical stage relative to the starting levels as we were in 2000:
The trajectory of the yield curve is following almost an identical path in the 12 months preceding the yield inversion:
If this scenario plays out again, I reckon it entirely plausible we have a fairly mild, normal economic recession but one in which it takes years for stocks to grind out to their eventual bottom. If this is a recession (and that is still very much a HUMONGOUS ‘if’), a recession might be expected to formally begin at any point from July 2019 to June 2020.
I’ll be watching the yield curve closely to see whether this is a mere fleeting blip (a la 1998) or the beginning of a more serious dive into inversion land.