It seems like everyone’s talking about the yield curve. Will it invert? If so, when? Would it imply or even precipitate a recession? How would it affect Fed policy? What would it mean for the stock market?

Since our work focuses on forecasting markets, let’s set aside for the moment whether an inversion means a recession is any more or less likely. After all, the stock market is not the economy. Instead, let’s turn our attention to how markets react to yield curve inversions.

February 2000

Since 2000, we’ve had four inversions: instances of 10-year yields falling below 2-year yields. The first began on Feb 2, 2000, about 7 weeks before SPX topped out and 5 weeks before COMP topped out. The inversion didn’t last long, however. By the end of December, the 2s10s had swung back to positive. SPX, which had recently corrected nearly 20%, rallied 10% by the end of January, peaking on the same day as the 2s10s (the white arrows on each chart.)

At that point, SPX began another leg down which would see it shed a total of 30% from the top, a third which totaled 40%, and a fourth which totaled 50%. Notably, these more dramatic plunges didn’t occur until 2s10s spiked higher.

When 2s10s broke out above a trend line (below, in red) which connected the former highs in 1996 and 1998, stocks’ losses accelerated. In other words, the initial inversion coincided with a modest 20% correction which lagged the 2s10s by 14 sessions.  It was the 2s10s breakout – a sharp rise from 33 bps to 234 bps once it cleared that red trend line – that coincided with the crash.

Once the 2s10s began its ascent, its oscillations continued to correlate with stocks’. Zooming in, we can see that the 2s10s repeatedly bounced off a rising yellow trend line (TL) for the next six months, finally topping out and dropping through the TL in June 2001 (point 1 on the chart below.) Initially, SPX consolidated.

When 2s10s popped up through overhead resistance (the purple TL) however, SPX began another sharp leg down. It happened again at point 2, breaking down through another yellow TL from the 2000 lows and popping up to new highs above the purple TL.

We can draw two general conclusions from this inversion:

1. First, the initial inversion is not a sell signal. If someone had gone to cash the day the 2s10s inverted, they would have missed the last 10% of the rally which followed;
2. While breakdowns often coincide with corrections, it’s the breakouts which coincide with crashes.

 

December 2005

Let’s look at the next inversion, which began in December 2005. The 2s10s bounced back and forth until June 2007 – five weeks before the legitimate July 16, 2007 high and a full 4 months before the October 11 high. Anyone rushing to cash out when the inversion first occurred would have missed out on the last 25% of upside.

Again, the correction began about the same time as the breakout. An 8% correction occurred in May 2006 (point 6) when 2s10s briefly turned positive. The big correction waited until shortly after the lasting breakout at point 8.

SPX shed 20% by the time 2s10s reached point 9, bounced 15% as 2s10s retraced to point 10, then cratered as 2s10s broke out of the recent trading range.

Note that the crash picked up steam when 2s10s pushed to new highs in Oct 2008.  When 2s10s topped out in November, SPX began bouncing again, rallying until 2s10s reached its former lows at point 12.

By March 2009, QE was taking hold. Stocks were rescued by massive amounts of liquidity which transformed the bond market from a barometer of economic conditions to a tool by which the Fed could signal algorithms to buy stocks.  It was the beginning of the end for price discovery.

Once the Fed assumed control of yields, all bets were off. The 2s10s rallied along with stocks and pretty much everything else that wasn’t nailed down.

The lessons learned from this inversion were quite similar to the one in 2000:

1. The initial inversion wasn’t a great sell signal – stocks rallied another 25% before finally topping;
2. While breakdowns often produce corrections, it’s the breakouts that produce the more serious crashes.

 

August 2019

We now turn our attention to the most recent period and the impending inversion everyone’s talking about.  Spoiler alert: There will be one, probably by Friday.

We’ll start with the big picture. Remember, the yellow TLs usually reflect horizontal levels through which 2s10s break down (and, sometimes, backtest and break out through.) The red TLs are longer term – typically multi-year – and highlight major breakouts which presage sharp rallies in 2s10s and corrections in stocks.

There has been no shortage of yellow trend lines over the past 12 years, with most breakdowns usually presaging a correction. Some examples:

– Point 15: 17%
– Point 16: 22%
– Point 18: 15%

But, of course, the real action occurred when 2s10s broke out above the red TLs.  The most glaring example is the one extending almost 6 years from December 2013 to October 2019. It validates the adage that “the bigger they are, the harder they fall” and set the stage for the 35% drop in Feb/Mar 2020 [see: Buckle Up.]

That breakout, however, was really a series of breakouts and breakdowns that helped position 2s10s for a very bearish move. A closeup illustrates, for instance, the 20% correction beginning in Sep 2018 which followed the breakout of a red TL from Feb 2018 (point 21) and the subsequent breakdown below horizontal support.

2s10s spent the next year in a very narrow range, finally dropping through horizontal support again and becoming inverted in August for a grand total of 2 days. It wasn’t exactly dramatic. In fact, SPX actually rallied in what would turn out to be the calm before the storm.

After the inversion, 2s10s push up through the yellow TLs representing previous support (now resistance) and three separate multiyear red TLs. When it finally cleared the last one (the white arrow) the equity market fell apart, crashing 35%.

I should mention here that it also pushed above a very important trend line we haven’t yet discussed, the dashed white TL connecting previous inversions dating back to 1995. I mention this because at point 25 we also broke back down through this same TL, all but guaranteeing that SPX’s recent bounce is over and lower lows are in the offing.

April 2022

Since topping in March 2021, the 2s10s has broken down through multiple yellow TLs of support. It also dropped through the dashed purple TL connecting the 2000 lows and the Sep 2019 lows and another TL from 1980. If you knew nothing else about investing, you would correctly assume that this positions it for the correction which has already begun.

The 2s10s has already inverted intraday a few times this week (also coming very close, therefore, to breaking below the Sep 2019 lows.) As we’ve seen in past, inversions aren’t in and of themselves terribly bearish for stocks.  It’s the bounce, subsequent rally, and spike higher following its inversion that do the most damage to stocks.

Let’s take just a moment to discuss what causes inversions. Below we can see a chart of the 2Y versus the 10Y for the past few years. Periods during which SPX corrected are shaded in blue. We can see that just prior to those periods, the 10Y and 2Y converge. Think of it as a coiling.

When stocks finally plunge, we see the 2Y drop much more rapidly than the 10Y, causing the gap between them to rapidly grow larger – the spike in the 2s10s.The same thing has happened almost every time since 2000. The single anomaly was in 2011, when the 10Y dropped much further and more rapidly than the 2Y.  Recall that this coincided with the US credit downgrade and it precipitated a 21% correction in SPX anyways. It also coincided with the most precise crash warning I have ever issued, but that’s another story.

As discussed above, the 2Y’s plunge hasn’t always occurred immediately after the convergence. In some cases, such as 2006, they continued in lockstep for over a year.

But, combined with the signals generated by the breakdowns and breakouts in the charts above, we have a very strong argument that the correction is not finished and could, in fact, get much worse.

If you’re thinking to yourself “sure it works in practice, but what about in theory?” there’s a very simple explanation for all of the above. Think about those previous equity meltdowns. As stocks rolled over, short-term treasuries yielding anywhere from 0-5% with little downside risk were very preferable to equities with plenty of downside risk.

Remember, markets aren’t about those participants in the tails changing their tune. Bears will be bears, and everyone else buys the dip.  But, when the 90% of all investors who are direction agnostic (index funds, program sellers, etc.) panic or are stopped out, it doesn’t take much of a shift for the pendulum to swing in favor of the bears.

There are also many fundamental economic arguments for the inversion and its unwinding. The most important one is that as the economy begins to slow, recession fears increase. Capital is reallocated. Risky long-term investments are reined in and short-term, fixed income investments are favored.

In the past, the Fed has “fixed” this phenomenon by reducing short-term rates dramatically enough to stimulate more risk taking. But that’s difficult to do when rates have spent years at zero, and impossible to do when said policy has led to an inflationary spiral that has the economy teetering on the verge of stagflation.

I’ll leave you with one last chart. For the past few decades, the 10Y has followed a very well-formed path. The channel shown below has been extremely reliable in forecasting tops. It was also very reliable in forecasting bottoms until 2020, when a return to the channel bottom would have meant negative interest rates.

The 10Y recently completed an Inverted Head & Shoulders pattern, a fairly reliable pattern that in this case targets 3.2%. Given that inflation is currently pushing 8%, a 3.2% 10Y would be, if anything, too low.

But, it would also cause the 10Y to break out of this 30-year channel – a very serious threat given that the national debt now exceeds $30 trillion. The Fed can’t be excited about letting this happen. They probably have this very same chart on the wall, smacking it for luck a la Ted Lasso every time they head into an FOMC meeting.

In any case, the channel top is currently at 2.58% – meaning the Fed is officially between a rock and a hard place. They have little choice but to watch the 30-year downtrend in rates to come to an inglorious end in an effort to rein in the inflation that they obviously and inevitably caused.

It’s simply too late to engineer an economic slowdown that would resolve the inflation problem without entailing higher rates. Given the Ukraine invasion and the shortage of journalists willing to give their lives in exchange for lower oil/gas prices, CPI will remain elevated for at least a few more months.

If stocks fall far and fast enough, the 2s10s could invert for more than a few days – a coiling that would ultimately unleash an even larger correction. My crystal ball says that the S&P 500 will drop to 3956 and possibly 3855 in the next month or two. If those support levels don’t hold, we could be in for a drop as low as 3047.

Should we fear a yield curve inversion? Not particularly. But what comes after one could ruin your whole day. Stay tuned.