It is true that much of the weakness in the nowcast is identified by economic variables that relate to the industrial sector. But these variables have, in the past, been very closely correlated with activity in the economy as a whole, and are therefore usually among the best indicators of overall activity. It is dangerous to ignore weakness in these industrial variables that persists for a long period, which is what is happening now....The full model, including the industrial sector data, estimates that the recession probability has been hovering around 15-20 per cent (above right graph), no longer an entirely negligible risk. If the weak industrial data are excluded, on the grounds that they are “transitory” – a word often used by Fed officials – then the recession probability drops to about 10 per cent.
He adds this picture:
Recession odds of just 10% would hardly be worth getting out of bed for. So how much weight should we be placing on the manufacturing data? I often see claims that manufacturing is already in recession. And Andrew Levin, former advisor to Federal Reserve Chair Janet Yellen, places much weight on the industrial production slowdown:
Unfortunately, the latest economic data underscore the risk that the economy may now be heading into another recession. Last Friday, the Federal Reserve Board reported that its index of industrial production sank further in December and was down 1.8% from a year earlier. Indeed, as shown in the accompanying chart, this pace of contraction has only occurred during prior recessionary periods. In some instances, the fall in industrial output was a harbinger at the onset of a recession. In other episodes, the industrial sector had been booming previously and turned downward after a recession was already underway. But since 1970 there has never been a case where the industrial sector shrank nearly 2 percent on a 12-month basis and the broader economy was left unscathed.
I think it is important to be very cautious with this aggregate data. What makes a recession a recession is that the decline in activity is felt widely throughout the economy. From the National Bureau of Economic Research:
During a recession, a significant decline in economic activity spreads across the economy and can last from a few months to more than a year.
With this in mind, I direct you to my fellow Oregon economist Josh Lehner, who correctly notes that in comparison to past recessions, the decline in manufacturing activity is not well disbursed across the sector. My version of Josh's chart:
The point is that during a recession, the vast majority of manufacturing industries (or all!) are declining. We are nowhere near that point. In other words, even manufacturing - arguably the most distressed sector of the US economy - is not recession. And if manufacturing is not even in recession, it is difficult to see that the US economy is in recession. Or even nearing it.
Initial unemployment claims across states tells a similar story:
In this version, I count the sates experiencing a 5% or greater change in year-over-year unemployment claims (I used 5% to account for the fact that as the cycle matures, claims will flatten out for more states and thus you would expect a wider dispersion of marginally higher claims). As is evident, recessions are characterized by rising claims across a wide swath of sates. In other words, a recession in Texas does not a US recession make. Note also that the economy can experience a fairly widespread increase in claims but not a recession. See 1995. Which means that while I think initial claims is an excellent leading indicator, it by itself is not infallible.
Aside from the recession risk, there is another important aspect of Davies's chart - discounting manufacturing, it indicates growth of just 2% in the US. This is fairly close with the Federal Reserve's estimate of potential growth, and I suspect that is the direction we will be heading by the end of the year if not sooner. Key sources of growth, such as autos, multifamily housing, and technology, that helped propel the economy closer to fully employment are likely leveling off. If so, that means the economy is at an inflection point as it transitions back to trend. The Fed expects that process will require addition tightening. The financial markets aren't so confident.
Bottom Line: The lack of widespread economic weakness across the economy indicates that the US is not currently in recession. It is not even evident manufacturing is in recession. If the economy were heading into recession, expect the dispersion of weakness will spread further across the economy, both geographically and sectorally.