In 2019, Fed economist David Miller undertook a comprehensive assessment of term spread predictive power for recessions (There is No Single Best Predictor of Recessions). For the 1984-2018 period, he found the following:

(Click on image to enlarge)

Figure 2: AUROC Sample 1984 – 2018 from Miller (2019).

For a discussion of AUROC (Area Under Receiver Operating Characteristics curve), see this post by Jim H.

What do these spreads look like as of Friday?

(Click on image to enlarge)

Figure 1: 1yr-Fed funds spread (chartreuse), 5yr-Fed funds (red), 10yr-Fed funds (bold teal), 10yr-1yr (sky blue), and 3yr-2yr (purple), all in %. Red dashed line at expanded Russian invasion of Ukraine. Source: Treasury via FRED, and author’s calculations.

The 1yr-Fed funds spread (1-2 month horizon) has not gone negative, nor has the 5yr-Fed funds spread (4-9 months) although it’s come close. The 10yr-Fed funds spread (10-19 months) did end negative on November 10th, which would suggest that somewhere between September 2023 and August 2024 a recession would occur. The 10yr-1yr has also registered negative on July 12th, so between 20 and 23 months ahead (February to May 2024).

Here by the way are the usual spreads I show:

(Click on image to enlarge)

Figure 2: 10yr-Fed funds (teal), 10yr-3mo (blue), and 10yr-2yr (dark red), all in %. Red dashed line at expanded Russian invasion of Ukraine. Source: Treasury via FRED, and author’s calculations.

Note that the AUROCs were calculated for simple term spreads, and not models incorporating other variables like foreign term speads, financial conditions indices or stock prices (as discussed in this post).

More By This Author: