We survey households about their expectations of the economic fallout of the COVID-19 pandemic, in real time and at daily frequency. Our baseline question asks about the expected impact on output and inflation over a one-year horizon. Starting on March 10, the median response suggests that the expected output loss is still moderate. This changes over the course of three weeks: At the end of March, the expected loss amounts to some 15 percent. Meanwhile, the pandemic is expected to raise inflation considerably. The uncertainty about these effects is very large. In the second part of the paper we feed the survey data into a New Keynesian business cycle model. Because the economic costs of the pandemic have not fully materialized yet but are nonetheless (a) anticipated and (b) uncertain, private expenditure collapses, thereby amplifying and bringing forward in time the economic costs of the pandemic. The short-run economic impact of the pandemic depends critically on whether monetary policy accommodates the drop in the natural rate of interest or not.

The notion that flexible exchange rates insulate a country from foreign shocks is well grounded in theory, from the classics (Meade, 1951;Friedman 1953), to the more recent open economy literature (Obstfeld and Rogoff, 2000). We confront it with new evidence from Europe. Specifically, we study how shocks that originate in the euro area spill over to its neighboring countries. We exploit the variation of the exchange rate regime across time and countries to assess whether the regime alters the spillovers: it does not -flexible exchange rates fail to provide insulation against euro area shocks. This result is robust across a number of specifications and holds up once we control for global financial conditions. We show that the workhorse open-economy model can account for the lack of insulation under a float, assuming that central banks respond to headline consumer price inflation. However, it remains puzzling that policy makers are ready to forego stabilization of economic activity to the extent we found in the data.

No abstract