As I mentioned before, I'm now more on recession watch than I am on inflation watch. As in the OP, I'm organizing my thoughts about the best data to get warnings about recessions, to obtain confirmation a recession is in progress, and to obtain confirmation a recession is ending.
I'd like to (1) get advance notice when a recession starts, (2) confirm without a doubt a recession is here, and also (3) get a sense of when the turning point occurs, which would be the ideal time to go long.
Of course, if one could get an accurate read on the start date and turning point of a recession, then one could enter a bear spread on a stock index prior to the recession, exit that spread for a big gain, and then enter a bull spread at the end of the recession, earning another big gain on the rebound. That would be a 200+% gain for two well-timed ATM spreads. A more conservative play would be to just ride the recession in cash/treasuries and get back into stocks when the recession is going badly, thus avoiding a major Sequence of Returns Risk Event. Call it a fool's errand if you want, but let's look at the sorts of data we have at our fingertips:
Advance Recession Signals10y/2y yield curveThis classic yield curve seems to offer the most advanced warning of a recession. This signal has enjoyed great predictive accuracy in the past, and offered investors ample lead time unlike many other signals. A brief and shallow false signal occurred in 1998 during the Asian financial crisis, but other than that the 10/2's record remains strong in the past half-century. A strong signal has always meant recession is on the way, but it can take over a year to arrive. The 10/2 inverted on July 6, 2022 and remains deeply inverted today. Today's -0.66% spread is the largest since 1981.
10y/3m yield curveThe 10y/3m curve is an even more reliable recession warning than the 10/2 with the last failure occurring in 1967 in a very different economy (gold standard, no dual mandate, no QT/QE). If investors are willing to accept lower yields to take on 10 years of duration risk rather than 3 months, that means they expect something bad to happen and rate cuts to arrive soon. It's an extreme inversion of the normal premium borrowers must pay to borrow for longer periods. Also note how both the 10y/2y and 10y/3m tend to un-invert a few months before the start of recession, providing a false dawn or calm before the tsunami. This indicator is warning of recession within the next year or so.
A 300+bp Federal Funds Rate IncreaseBy the beginning of February, the FFR will have gone from a higher bound of 0.25% to 4.75%. There has not been a period in modern economic history where rates rose by at least 450bp within a couple of years and a recession failed to follow. 2020 followed a roughly +225bp campaign. 2008 followed a roughly +425bp campaign. 2001 followed a roughly +189bp increase. Rate hikes must be sufficiently large to cause a recession. The threshold of uncertainty appears to be somewhere just above 300bp. At or below that level of rate hikes, it appears the economy is capable of surviving without a recession, but above it abandon all hope! Between 1993 and 1995 there was a roughly 3% rate hike and a recession did not follow. Similarly, between 1983 and 1984 rates increased by about 3% and nothing happened. Rate hikes often precede recessions by a year or more. The last 4 recessions began after the Fed already started cutting rates, so simply following the Fed and going long as the rate cuts start is not a viable strategy. According to this indicator, the odds of a recession this year or next are nearly certain.
Conference Board Leading Economic Indicators (LEI) indexThis composite index offered advance warning of the 2001 and 2007-2008 recessions, and quickly signaled recession when the pandemic hit in 2020. It appears to offer a couple months of advance notice when a once-in-a-century pandemic-scale crisis is not occurring. There also seems to be a reliable "recession-is-over" signal because the coast seems to be clear when the LEI returns to zero percent. Currently the LEI index is signalling recession because it has fallen below the Conference Board's "recession warning" threshold.
National Financial Conditions IndexBanks and investors tend to tighten their lending standards when a recession is around the corner. The NFCI purports to measure the ease or difficulty of obtaining financing by combining a variety of spreads and swaption volatility into an index. When the NFCI rises above zero (it is normally negative) a recession is on the way. The NFCI is one of the more reliable recession predictors, but the timing between the signal and the recession's start is highly variable. Even worse, the NFCI made several false recession calls in the 80s and then missed the 2001 recession completely. Nonetheless, there's a direct conceptual link between tighter financial conditions and negative growth so we should pay attention.
Collapsing Natural Gas PricesNatural gas has historically run up ahead of recessions and then crashed just as the recession took hold. That was the case in 2001 and 2007-2008. However, gas prices also peaked in 2003, 2004, and 2005 with no associated recession. Currently, gas prices peaked in August 2022 and are now plummeting. It's probably too late to prepare for recession by shorting UNG.
----------------------------------
Real Time Recession SignalsInitial claimsWhen people were claiming that two consecutive quarters of negative GDP growth in 2022 meant a recession had occurred, I said "it's not a recession if nobody loses their job." It wasn't a recession either, because job growth actually occurred! To identify the start of this next recession, I'm looking for an increase in initial claims from last week's 190k/week (a historically low-water mark) to 330k to 350k (numbers reached just before the GFC, 2001 recession, and 1991 recession). If weekly initial claims rise to those levels, it means the recession is almost here. Today's super-low initial claims number suggests the recession might be a long time coming.
Private Residential Fixed InvestmentDips in residential investment have preceded most recessions since at least the late 1940s. The pattern has been for a decline in housing investment to precede a recession by a quarter or two, although there have been exceptions. When residential investment falls by about 5% or more, a recession is imminent or in progress. The combination of lower reliability (than say, yield curves) and the information arriving almost too late to make a difference (it's part of the GDP report) make this more of a recession confirmation metric than a predictor. Residential investment tends to recover in the second half of a recession.
CNN Fear & Greed IndicatorThis is an aggregate of stock price momentum & breadth, 52 week highs vs lows, puts vs calls, volatility, last 20 day stock vs. bond returns, and junk bond spreads. The resulting indicator is very noisy, but provides context to investor activity. During a recession, the
best only times to buy would be "extreme fear" times.
University of Michigan Consumer Sentiment SurveyLook at the 50-year chart of the Consumer Sentiment Survey and notice how sentiment plunges at the start of a recession. Also note that sentiment does not tend to improve until the recession is basically over. Thus, I suggest using consumer sentiment as a tool to identify the start of a recession. Today's sentiment has fallen to levels last seen during the GFC, but appears to be staging a recovery. So maybe the recession is close? This diagnosis would conflict with the patterns usually seen with yield curves.
-----------------------------
Recession Exit SignalsUniversity of Michigan Current Economic Conditions SurveyUnlike the U of M's consumer sentiment survey, the economic conditions survey seems to be an excellent indicator of when the end of a recession is imminent. See the 50 year chart. Consumers may be grumpy (sentiment) but they have a collective finger on the pulse of the economy and seem to know when economic conditions start to turn around. Of course, some credibility is lost when we note that consumers think the economy is worse today than it was during the GFC or the early 1980s, and yet unemployment is now only 3.5%. Jeez, what would they think if the economy was actually bad?
The NFCI, again.Note how the NFCI typically plummets at some point in the recession, and by the time it hits zero again the recession is over. This provides an end-of-recession signal that arrives faster than the NBER recession declaration.
Conference Board Leading Economic Indicators (LEI) index, againThe LEI has historically leapt back up toward zero, arriving at zero sometime after the recession has ended. This provides an end-of-recession signal that arrives faster than the NBER recession declaration.