An important disconnect in the news view of fluctuations is the lack of consistent evidence suggestive of significant macroeconomic effects of news shocks. Findings from estimated DSGE models that in theory allow news shocks to matter quantitatively suggest that they do not. This disconnect can be resolved once we augment a DSGE model with a financial channel that provides amplification to news shocks. Our results suggest that news shocks to the future growth prospects of the economy are significant drivers of U.S. fluctuations, explaining as much as 50% and 37% of the variance in hours worked and output, respectively, in cyclical frequencies.

We construct a simple model with lumpy investment, cash accumulation, and costly external finance. Based on this model, we propose a new savings specification aimed at examining savings behavior in the presence of investment lumpiness and financial constraints. We then test a key prediction of our model, namely that under costly external finance, savings–cash flow sensitivities vary significantly by investment regime. We make use of a panel of firms from transition and developed economies to estimate the new savings regression which controls for investment spikes and periods of inactivity. Our findings confirm the validity of the model’s prediction.