Conflicting signals for coinciding macro indicators at the end of May

Monthly indicators of employment, consumption, personal income (ex-transfers) are growing in April. But GDO and GDP+ show declines in Q4 2022 and Q1 2023.

Picture 1: Non-farm employment, NFP (dark blue), Bloomberg consensus of 5/26 (blue+), civil sector employment (orange), industrial production (red), personal income excluding transfers in 2012, $ (green ), production and trade sales in Ch.2012$ (black), consumption in Ch.2012$ (light blue) and monthly GDP in Ch.2012$ (pink), GDP (blue bars), all logs normalized to 2021M11= 0. The Bloomberg consensus level is calculated by adding the forecast change to the previous unrevised employment level available at the time of the forecast. Source: BLS, Federal Reserve, BEA 2nd Edition 2023Q1 via FRED. S&P Global/IHS Markit (nee Macroeconomic Advisors, IHS Markit) (5/1/ 2023 issue) and the author’s calculations.

Given the focus of the NBER Business Cycle Dating Committee on employment and personal income, one can be fairly confident that there has been no recession as of April 2023, bearing in mind, of course, that all of these figures will be revised over time. GDP, in particular, will be revised many times, so an increase in this series will not be critical to rule out a recession (in the same way that a decline in Q1-2 2022 will not be decisive to determine a recession).

We know that published GDP is not really the best indicator of the extent to which GDP will eventually be revised. GDO and GDP+ are the two series most likely to satisfy this condition. Here we see some worrying signs.

Figure 2: GDP (black), GDO (blue), GDP+ scaled to Q4 2019 (yellow), potential GDP (gray line), GDPNow 5/26 (red square), SPGMI 5/26 tracking (sky blue triangle) in billion hl. 2012 SAAR. Source: BEA, Philadelphia FedCBO (February 2023) Federal Reserve Bank of AtlantaS&P Global Market Insights and author’s calculations.

While GDP has been revised to 1.3% SAAR, GDO (average of GDP and GDI) and GDP+ are -0.5% and -1.2% SAAR respectively. Ass Jason Furman noted that the discrepancy between GDP and GNI is very large, highlighting the uncertainty we face in determining trends in economic activity. This shows up in a discrepancy in the bean counting exercises: GDPNow is 1.9% SAAR, while SPGMI (formerly Macroeconomic Advisors and IHS Markit) is 0.4% – practically zero.

As Furman points out, if GDP and VRT were the only series we observed, we would turn to VRT. But, again, we have a lot of evidence about the strength of the labor market. One of the bottom line indicators is the Philadelphia Fed’s coincident index for the US. In Figure 3, I show an overlapping index compared to non-farm employment and consumption.

Figure 3: Concurrency Index (Chartreuse), Nonfarm Employment (Blue), Bloomberg 5/26 Consensus (Blue+), Consumption (Sky Blue), All in Journals, 2021M11 = 0. Source: Philadelphia Fed, Bloomberg, BLS and BEA via FRED and author’s calculations.

The coinciding index is based largely on labor market performance and has steadily increased, even during the first half of 2022, when some observers argued that a recession had arrived. Consumption, which is mainly supported by wages and wage payments, has also risen significantly over the past year and a half and rose unexpectedly in April.

In general, uncertainty reigns!