Deputy Asst Secretary Tara Sinclair and Asst Sec Van Nostrand and Special Asst Gupta discuss the ongoing boom in manufacturing construction. Here’s one graph:

Source: Van Nostrand, Sinclair, Gupta (2023).

One doesn’t see movements like this in response to a non-macro policy. I wondered what this was in the context of overall private nonresidential fixed investment. here’s the picture. Instead of the construction data, I use BEA fixed investment data (which is a lot less than construction data for 2023Q1 — see appendix comparing the numbers).

Figure 1: Nonresidential fixed investment (bold black line), GDPNow for Q2 (red square), nonresidential fixed investment in manucturing structures (blue bar), in non-manufacturing structures (brown bar), and rest-of-nonresidential fixed investment (green bar), all in billions Ch.2012$ SAAR. Source: BEA, Atlanta Fed, and author’s calculations.

Looking forward, as of today, the Atlanta Fed’s GDPNow indicates the following for Q2 GDP and nonresidential fixed investment.

Figure 2: GDP (black, left log scale), GDPNow for Q2 (red square, left log scale), and nonresidential fixed investment (blue, right log scale), GDPNow implied nonresidential fixed investment (light blue square, right log scale), all in billions Ch.2012$ SAAR. NBER defined peak-to-trough recession dates shaded gray. Source: BEA, Atlanta Fed, NBER, and author’s calculations.

Even as GDP slows, investment is continuing to grow more rapidly, so that the log ratio of investment-to-GDP keeps on rising. GDPNow implies that continues into Q2. In the period of the Great Moderation, investment-to-GDP peaks before the NBER peak (aside from the Great Recession). This suggests that an onset of a recession is still some time off.

Appendix:

Here is the nominal version of the series used in the Treasury document, and the BEA series for investment in structures.

Figure 3: Construction in manufacturing (blue), and investment in nonresidential manufacturing structures (black), in billions $, SAAR. 2023Q2 observation is average of April and May data. NBER defined peak-to-trough recession dates shaded gray. Source: Census, BEA, and author’s calculations.

Brandsaas et al. (2023) argue that the Census series is likely to end up being more accurate, given that it is used to revise the BEA series.

More By This Author: