As of mid-December, it's hard to see a recession (even with GDP growth downgraded to 4.9% this morning), so there's a lot of triumphant sentiment about what the oft-predicted 2023 recession will look like. We only have the best monthly data through November, so maybe it's too early to declare 2023 a winner – but just a little bit. For 2024, these concurrency metrics are of little help.
First, the NBER Business Cycle Dating Committee focuses on the series as well as the SPGMI monthly GDP.
figure 1: Nonfarm payrolls included in preliminary benchmark (bold dark blue), using Bloomberg consensus implied levels as of 12/4 (blue+), civilian employment (orange), industrial production (red), 2017 excluding Transferred personal income $ (bold green), manufacturing and trade sales $ in 2017 (black), consumption $ in 2017 (light blue), and monthly GDP $ in 2017 (pink), GDP, third release ( Blue bar), all logs are normalized to 2021M11=0. Source: BLS via FRED, U.S. Bureau of Labor Statistics preliminary benchmarksFederal Reserve, BEA 2023Q3 2nd Edition, contains comprehensive revisions, S&P Global/IHS Markit (Nigerian macroeconomic consultant, IHS Markit) (12/1/2023 release) and the author's calculations.
Not shown in Figure 1 is the Atlanta Fed's GDPNow (12/19), which shows quarterly seasonally adjusted rate (SAAR) growth of 2.7% in Q4 2023.
Here are other popular indicators: Philadelphia Fed Coincidence Index, GDO, and Heavy Truck Sales. I also added some special measures such as vehicle mileage and gas consumption.
figure 2: Top panel coincidence index (blue), vehicle mileage (red), gasoline consumption (cyan), 2017 GDO$ (blue bar), all logarithmic, 2021M11=0; seasonal using X13 (ARIMA X11) Adjusted gasoline consumption. Bottom panel Heavy Truck Sales (light blue), centered on 3-month moving average (bold dark blue), both logarithmic, November 2021 = 0. Source: Federal Reserve Bank of Philadelphia, Department of Transportation, Census Bureau, BEA GDP 3rd Release (via FRED, EIA, and author's calculations).
The coincident index has grown much faster than the GDO, reflecting that it is mainly based on labor market indicators. Two travel metrics, VMT and gasoline consumption, appear to be particularly unreliable given working from home, and for gasoline consumption the growing popularity of electric vehicles. Heavy truck sales are still growing compared to the first half of 2021, but it is worth noting that sales are declining rapidly (after peaking in April 2023), which is a harbinger of an economic slowdown.
These indicators lag behind three months ago. Here are some weekly economic activity indicators (tracking year-on-year GDP growth) released today as of a few days ago.
image 3: Lewis-Mertens-Stock Weekly Economic Index (blue) and Baumeister-Leiva-Leon-Sims US Weekly Economic Conditions Index plus 2% Trend (green), all annual growth rates in %.Source: Federal Reserve Bank of New York fred, WECIaccessed 12/21, and author's calculations.
The WECI +2% as of 12/16 was 1.73%, while the WEI reading was 2.36%. If the 2.36% reading persists throughout the quarter, the latter could be interpreted as a year-over-year increase of 2.36%.this Baumeister et al. A reading of -0.27% is interpreted as growth below the long-term trend growth rate of 0.27%. The average US GDP growth rate during 2000-19 was about 2%, so this implies a growth rate of 1.73% as of 2012/16.
So we have some mixed signals about year-over-year growth in the economic data released in mid-December.
Recall that the WEI relies on ten series of correlations for which weekly frequency is available (e.g., unemployment claims, fuel sales, retail sales), whereas the WECI relies on a mixed-frequency dynamic factor model.
