Beaudry and Portier (2006) provide support for the "news view" of the business cycle, using a vector error correction model. We show that this result hinges on a cointegrating relationship between TFP and stock prices that is not stationary, thus making the estimates not reliable. If we alter the TFP measure and change the model specification, we can recover the news shock through their identification. However, the news shock leads to a stock market boom with a negligible impact on economic activity. Our findings are in line with studies that identify news shocks without relying on VEC models.
cointegration technology news shocks stock prices TFP VEC model SVAR