There is no obvious recession as of August 2024 data (my interpretation – see opposing views here). Using last year’s financial data, what would the steepest decline indicate? It matters which variable you use, and whether you include the 2020 recession.
Figure 1: Next year probit regressions are estimated, using the time spread of 10yr-3mo and the short term (bright blue), the same using the final sample before the pandemic (blue), the time spread, the short term and the spread of the external term (tan), term-premium adjusted 10yr-3mo term spread and term premium and debt-service ratio (green). Sample 1986-2024M08, assuming no recession started in 2024M08. The NBER has defined recession days as shaded in gray. Numbers show estimated probabilities for September 2024. Source: author’s calculations.
Discussion on the use of the foreign name, a la Ahmed-Chinn (JMCB, 2024) here; and the use of the debt service ratio a la Chinn-Ferrara (2024) here. Skipping the word premium discussed here.
Note that the probability associated with the term spread/short-term specification is highly dependent on the sample used, especially whether the 2020 recession is included in the term spread and short-term specification. Using distribution data only for the term up to 2018M12 (so recession/no recession up to 2019M12) and forecasting out of sample yields a 78% chance of recession in September (this month). Incorporating a 2020 recession into the sample results with only a 48% probability this month (58% high average in May). If one thinks that the recession in 2020 would not have happened without this pandemic, then one can rightly argue with the release of that data.
So… a recession is still possible (or it’s already here and we don’t know it).
Source link