and UK GDP fell for two consecutive quarters Based on current data, it is reasonable to ask whether the UK is in recession. ONS discusses the limitations of using the two-quarters rule of thumb here. This picture 2 study illustrates the dangers of relying on the two-quarter rule when GDP data need revision. See other discussions using the definition of “technology”, Associated Press, CNN, Bloomberg. It is also interesting to consider whether statistical methods relying on financial indicators can predict recessions of different definitions.
Here is a picture before GDP, monthly GDP and industrial production:
figure 1: Real GDP (blue bar, left logarithmic scale), monthly real GDP (black line, right logarithmic scale), pre-construction industrial production (red, right logarithmic scale). Source: National Bureau of Statistics. The hypothetical June to December 2023 recession peak to trough appears in light blue.
I put the end of the recession in February 2023 because I have no other data beyond that time, and Bank of England Governor Bailey thinks recession may be over.
As the opposite, Nice SR observed:
…The state of the UK economy is better described by the fact that GDP fell between the first quarter of 2022 and the last quarter of 2023.
Therefore, the recession may be from the second quarter of 2022 to the fourth quarter of 2023, from peak to trough.
Would we have predicted either of these recessions (if one was indeed a recession)?I compare the estimate of the simple 10-Year-March spread plus short-term rates (blue line) to the estimate of the foreign term spread plus debt service ratio (brown line), which is given by Chin-Ferrara(2024).
The answer is probably no (for the 1985-2022 sample) unless a very low threshold is used.
figure 2: Probit model estimated probabilities for interest rate spreads and short-term rates (blue line) and foreign interest rate spreads and debt service ratios (brown line). ECRI defines the peak to trough date of a recession as a lavender shade. The hypothetical June-December 2023 recession is colored light blue. Source: Author's calculations and ECRI.
Although the regression specification based on the debt service model has a high pseudo-R2 (0.38), the implied recession probability in 2023 is quite low. Interestingly, the pseudo-R2 of spreads plus short rates is very low (0.08) but is associated with a much higher probability of recession – albeit well below 50%. This contrasts with research showing that interest rate spreads are good predictors of recession (Mills, Capehart, & Goodhart, 2019), even though their samples are much longer than the ones I used.
