Forecasting the End of the Global Recession : Did We Miss the Early Signs?

This paper looks at the term-structure literature to identify early signs predicting recessionary patterns in the U.S. and other developed economies. Based on the NBER and ECRI recession dates, we define the probability of recession as a function of the traditional yield spread plus a forward-looking measure of growth expectations, namely the output gap growth spread. For other countries, we extend the model and make it additionally dependent on the probability of recession in the U.S. Our results indicate that most of the a-posteriori official recession dates could have been forecast as early as April 2009, when the first green shoots appeared in the U.S. Overall, the term structure versions we apply allowed us to signal recessions earlier and presumably more accurately than traditional term structure models and most professional forecasters