Working Paper: CEPR ID: DP13519
Authors: Charles A. Goodhart; Terence C. Mills; Forrest Capie
Abstract: This paper investigates whether the inversion of the yield spread, with short-term rates higher than the long-term rate, has been and remains an effective predictor of recessions in the U.K. using monthly data from 1822 to 2016. Indicators of recession are constructed in a variety of ways depending on the availability and properties of the data in the pre-World War 1, inter-war, and post-World War 2 periods. It is found that, using peak-to-trough recession indicators and a probit regression model, there is reasonably strong evidence to support the inverted yield spread being a predictor of recessions for lead times up to eighteen months in all three periods.
Keywords: yield spread; recession; prediction; probit models
JEL Codes: E30; E32; E43; E44; N10
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| inverted yield spread (E43) | recessions (E32) |
| inverted yield spread (pre-World War I) (N13) | recessions (E32) |
| inverted yield spread (interwar years) (N13) | recessions (E32) |
| inverted yield spread (post-World War II) (N22) | recessions (E32) |
| alternative recession measure (cyc0) (E32) | inverted yield spread (E43) |