We revisit the spillover effects of US monetary policy using a new data source, the daily nighttime light (NTL), as a high-frequency proxy for real economic activities. Taking China as an example, the unexpected US tightening dampens its output and peaks about seven weeks later. This is explained by a construction channel, with the NTL variation mainly driven by non-built-up areas instead of city centers and suburbs. Consistently, construction-related indicators are harmed and cities with more construction activities and tighter financial conditions respond more negatively. Moreover, the US shock also affects output in other economies, especially emerging countries or regions.

scoco@ceibs.edu