In 2019, Fed economic expert David Miller carried out a detailed evaluation of term spread predictive power for economic downturns (There is No Single Best Predictor of Recessions). For the 1984-2018 period, he found the following:
Figure 2: AUROC Sample 1984– 2018 from Miller (2019 ).
For a conversation of AUROC (Area Under Receiver Operating Characteristics curve), see this post by Jim H.
What do these spreads look like since Friday (color coordinated to match the chart above?):.
Figure 1: 1yr-Fed funds spread out (chartreuse), 5yr-Fed funds (red), 10yr-Fed funds (vibrant teal), 10yr-1yr (sky blue), and 3yr-2yr (purple), all in %. Red rushed line at expanded Russian intrusion of Ukraine. Source: Treasury via FRED, and authors calculations.
The 1yr-Fed funds spread out (1-2 month horizon) has not gone negative, although the 5yr-Fed funds spread (4-9 months) has because the last post on this topic a month back. The 10yr-Fed funds spread out (10-19 months) did end negative on November 10th, which would suggest that somewhere in between September 2023 and August 2024 an economic downturn would happen. The 10yr-1yr has also signed up negative on July 12th, so between 20 and 23 months ahead (February to May 2024).
Keep in mind that these are maximum AUROCs for easy term spread designs. They do not incorporate foreign term spreads (as recommended by Ahmed and Chinn, 2022), nor an aspect based on disaggregated sectoral dividend yield ratios (as suggested by Chatelais, Stalla-Bourdillon, and Chinn, 2022)
.
