Dr. Rosser asks what is the evidence for term spreads predicting recessions, with special reference to Japan. Here is my partial answer:
Chinn and Kucko (2015) examine data over the 1970-2013 period, and obtain the following results:
Source: Chinn and Kucko (2015).
The Economist article on this subject is here.
Note that this specification includes a three month yield in addition to the 10yr-3mo spread. The results are largely the same for the simpler specification using only the spread, except that the spread is then not statistically significant for Japan.
There are surely numerous other studies of how the term spread predicts recessions or growth, for sets of countries; Haubrich (2020) reviews some of those studies. However, the more recent studies usually rely on panel analysis. See for instance Gebka and Wohar (2018) or Borio, Drehman and Xia (2018), or Hasse and Laujaunie (2020), so we don’t know what specification does well for a given country. (And of course, Mehl’s (2009) study is the most comprehensive for emerging/developing countries.)
My cautionary tale for China, here.
Ahmed and Chinn (2022) discusses the 10yr-3mo term spread, augmented by foreign term spread, as a predictor.