Working Paper: NBER ID: w29300
Authors: David G. Blanchflower; Alex Bryson
Abstract: We examine the start date of the Great Recession across OECD countries based on two successive quarters of negative GDP growth recession. For most OECD countries this establishes the start of recession in Q22008 or Q32008. We find that the Sahm Rule identifies the start of recession in the US to the beginning of 2008 but in other OECD countries it identifies the start in almost every case, after that identified by GDP. But the GDP and labor market data are subject to major revisions, so the turn is not apparent in most countries for some time. We establish our own rule for predicting recession using the fear of unemployment series to predict recession. It involves looking for a ten-point rise in the series compared to its previous twelve month low. These surveys are timely and have the major advantage they are not subject to revision. Across the OECD we confirm this finding with other types of qualitative data and especially so in the UK. Qualitative surveys, we show, in the US in 2006 and 2007 predicted the subsequent recession and they did the same in Europe at the end of 2007 and in the early part of 2008.
Keywords: Great Recession; Sahm Rule; OECD; Unemployment; Recession Prediction
JEL Codes: E17; J60; J64
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Sahm rule (C20) | recession onset in the US (F44) |
| Sahm rule (C20) | earlier recession identification than GDP data (E01) |
| fear of unemployment rise (J64) | recession prediction (E37) |
| qualitative surveys (C83) | recession prediction (E37) |
| GDP revisions (E01) | complications in recession timing identification (E32) |
| labor market loosening (J49) | recession indication before GDP decline (E20) |