exist Chin and Ferrara (2024), we find significant explanatory power for 10yr-3mo, 3-month interest rates, and debt service ratios in the German recession (our analysis also covers Canada, France, Italy, Japan, Sweden, and the United Kingdom). Our sample extends to the first month of 2022 (using spreads as of the first month of 2021) because we are uncertain whether a recession will occur in 2023. Since Germany is likely to experience an economic slowdown, it is necessary to consider the predictions of our model.
figure 1: GDP in constant euros (blue), industrial production excluding construction (blue line), employment (red line), all expressed in logarithms, 2022M09=0. Source: Eurostat, MEI (via FRED) and author's calculations.
The last peak of GDP will be in the third quarter of 2022. Industrial production peaks in February 2023, while employment continues to rise, pausing in January 2023. I tentatively set a date for the onset of the recession.It is important to note that neither German Council of Economic Experts, the ECRI didn't establish a peak either, so that's my assumption. (Laurent Ferrara provided a chronology up to 2019, as noted here).
Suppose I estimate the probability regression (using spreads, short-term interest rates (i3mo) and debt ratio (Digital SR) to 2020M12) to estimate the probability of economic recession in the next year. Guess so:
Pr(Record=1) = -4.85 – 60.6 spread +5.7 i3mo + 33.9 Digital SR
Pseudo R2 = 0.32, Nobs = 430, Smpl is spreadWait for 85M01-20M12. black body Expresses significance at 10% MSL.
The estimate of recession probability is shown in black, the ECRI-defined recession date is in light blue, and I assume the last peak is July 2023.
figure 2: Recession probabilities estimated using probabilistic models through December 2021 (black) and January 2024 (tan). ECRI-defined recession peak-to-trough dates are shown in light blue. Source: ECRI, author's calculations.
Note that the regression ignores the 2020 pandemic recession, although the odds do rise from very low levels. The chance of a recession in January 2024 is 19%.
Suppose I use the entire sample period up to January 2024 (and spreads and other determinants up to January 2023). Then we get the following estimate:
Pr(Record=1) = -4.05 – 59.7 spread + 6.54 i3mo + 27.8 Digital SR
Pseudo R2 = 0.29, Nobs = 455, Smpl is spreadWait for 85M01-23M01. Bold indicates significance at 10% msl.
The estimated odds for this regression are roughly the same, but the January 2024 odds are slightly higher at 23%.
In both cases, the model seemed unable to predict what would happen when a recession might begin. Having said that, no one has confirmed that Germany or the Eurozone are in ongoing recession.
Note that different results are produced if different chronologies are used, two of which (from Ferrara(2019)and General Certificate of Secondary Education (2022))As shown below.
notes: Peak to trough for all recession dates are shown in light blue. source: European Institute of Chemistry, Ferrara(2019), Germany Central and Eastern Europe (2022).
Well, the answers to these two questions are (1) maybe, and (2) maybe.
