2 weeks prior are used as instrumental variables for causal inference. Stringent quarantines, city lockdowns, and local public health measures imposed in late January significantly