Following the recent financial crisis and its subsequent Great Recession, the issue of a sluggish US employment was raised by economic observers. In a previous post on Econbrowser, Menzie Chinn pointed out the usefulness of the Okun’s law in assessing the potential level of employment after the recession. Especially, Menzie shows that:
- If one does not account for the long-term relationship between GDP and employment (i.e.; if one focuses only on the relationship in differences), then the bounce-back in employment after the 2008-09 recession cannot be captured.
- A standard error-correction model (ECM) is able to reproduce the general evolutions, but misses a large part of the recovery after the end of the recession.
- Accounting for the US business cycle by incorporating a dummy variable that takes the value 1 during recessions and 0 otherwise, according to the NBER Dating Committee dating, enables a better reproduction of stylized facts.
In this work, we reconsider this ECM approach but we do not impose any dummy variable and propose to let the data speak through a non-linear ECM. Read more