BofAML is “going there”.
“Yield-curve inversion is no longer a reliable signal for recession”, the bank’s Ethan Harris and Aditya Bhave write, in a note dated Friday.
The “is this time different?” question has been on the tip of various tongues since December, when inversions in the 2s5s and 3s5s kicked off a veritable panic which, by the end of the month, had large swaths of the investing public convinced that a US recession was imminent despite the fact that the underlying data suggested such an outcome was (basically) a mathematical impossibility.
We spilled gallons of digital ink attempting to refute the “imminent US recession” story late last year. If you haven’t read the pieces linked below, you should.
To be sure, the outlook for the global economy hasn’t materially improved – or at least it hadn’t until the very latest activity data out of China suggested an inflection point may be near. Warnings from the EC, the IMF, the WTO, etc. have all underscored the slowdown narrative and while the US continues to hold up, there is no sense in which we’re “out of the woods”, so to speak. (Far be it from us to whistle past the graveyard)
Meanwhile, the inversion of the 3M-10Y curve stateside (last month), reinvigorated the US recession story and reams have since been written about the history of inversions, lead/lag times to downturns and the performance of various assets.
But the world is different now, and if you ask BofA, the correlation between the 2s10s and recessions cannot be mistaken for causation. Critically, the bank notes that the Fed is “not an innocent bystander when the curve inverts: it can prevent or quickly undo inversions that it does not want.” Inversions, the bank writes, “are not accidents”.
Given that, it makes little sense to insist that the Fed is “sleep-walking over a cliff”, as many a commentator has asserted and as many a reporter/blogger has written in a headline. “The Fed is well aware of the historical correlation between yield curve inversions and recessions”, BofAML notes, before reminding you that there’s a sense in which trotting out the tired, old “they’re not getting the memo” line is entirely illogical. To wit:
One of the early influential studies on this issue was done by economists at the New York Fed. Research on the topic has intensified in the last two years, with many papers from around the Fed system, including two from the Board of Governors and four out of the San Francisco Fed. And scrolling through speeches, virtually every FOMC member has commented on the yield curve. The Fed did not just get the memo on curve inversion, it has basically been writing the memo.
Additionally, the Fed obviously doesn’t “want” a recession. Contrary to what Larry Kudlow, Stephen Moore and Donald Trump might tell you, the Fed is not a gang of growth-hating masochists hell-bent on intentionally undercutting the economy for the sheer sport of it.
Why, then, is the Fed seemingly countenancing inversions? For BofA, the answer is pretty simple:
It is not sleep-walking over a cliff; instead, after careful analysis the Fed has concluded that an inverted curve is not what it used to be and never was the sole relevant financial indicator.
The bank cites the familiar line of reasoning which says that with global term premiums structurally depressed, inversions are likely to happen far more often and then, Harris rolls out the “observer effect”. Here’s the key bit:
This is the theory that simply observing a situation or phenomenon necessarily changes that phenomenon. This is often the result of instruments that, by necessity, alter the state of what they measure in some manner. How does this apply here? Before Fed research came out, it was reasonable to argue that the Fed was not fully aware of the implied recession dangers. However, now the Fed has studied the relationship it will prevent or quickly undo inversions that it does not want.
Finally, Harris and Bhave touch on something that at least a couple of desks have mentioned this year with regard to the “good news is bad news” narrative, wherein upside data surprises can be negative for risk assets if they open the door to more hawkish monetary policy.
As it turns out, if you try to observe that phenomenon in the wild, the story doesn’t hold up particularly well, at least not on discrete days. “The markets price in Fed hikes (rates rise but the curve flattens) on days when data surprises are the most positive, and price out the Fed (rates fall but the curve steepens) in response to the most disappointing data releases”, BofA says, recapping.
But, if you look at days when the market prices in Fed hikes on good data, equities rally on average, while stocks sell off when data disappoints and the curve steepens as a result of rates traders pricing out the Fed. “In other words, a flatter curve is generally good news if it is accompanied by higher rates and a steeper curve is generally bad news if it comes with lower rates”, Harris says, summing up.
We would caution that a more exhaustive study of the “good news can be bad news in the post-crisis monetary policy regime” theory is probably in order before anyone draws definitive conclusions. While it may be true that, as Harris says, days when “bad news is good news” (to the extent disappointing data leads to risk asset rallies on the assumption that souring econ tips a more accommodative Fed) “are more the exception than the norm”, we can all think of countless instances where disappointing morning data releases in the US were met with pops in S&P futs on the assumption that the poor data took some pressure off the Fed to hike rates or otherwise stick to its normalization guns.
Additionally, we would reiterate what we said last Sunday, which is that currently, this is largely irrelevant. To wit, from last week’s week-ahead preview post:
If you had any worries that market participants might soon find themselves faced with a familiar post-crisis paradox where “good” news can be “bad” news to the extent it presages less accommodative central banks, you can probably put those worries to bed – at least for the next couple of months.
It’s now abundantly clear that policymakers have no stomach for a further deceleration in global growth amid increasingly dire warnings from the likes of the IMF, the WTO and as the picture continues to darken in Germany, where the government is set to slash its growth target in 2019 to just 0.5%, below the already lowered estimates of the country’s leading economic institutes.
With inflation still (forever?) subdued and as Donald Trump continues to pressure the Fed to ease policy, the bar for a hawkish turn from the world’s major central banks is arguably so high as to be unclearable in the near-term.
That sets the stage for any good news to be interpreted “purely” on the merits.
For now, we’ll just leave you with some excerpts from a conversation we had back in January with a strategist who weighed in on the perils of confusing correlation with causation when it comes to the curve and recessions:
There’s an issue of causality when you say inversion ’causes’ recession or recession ’causes’ stock market declines. This is not always correct and can be misleading. 1999 was an internet bubble burst [and] not much to do with Fed hikes; it just ran its course and died (of ‘natural’ causes). During that time, productivity growth was >3% and the economy was calibrated accordingly. When the bubble burst there was recession and Fed had to cut rates.
2004-7 was different. The securitization industry prevented transmission of rate hikes to the consumer. Although Fed hiked 400bp (they had to because inflation was rising), consumers were borrowing low through arms, so the Fed had to overshoot and the curve inverted. But that by itself didn’t cause the recession. The recession was triggered when arms started resetting and subprime borrowers felt all 400bp at once and started the chain reaction of defaults. This recession was a direct consequence of the Fed, but not Fed hikes as much as Greenspan‘s incompetence. The curve steepened in anticipation of rate cuts and equities declined, because it was the end of the housing bubble. And again, the economy was calibrated for it – productivity growth was >3% etc. But one cannot say that the steepener was an indicator of imminent equities decline.