We solve a canonical, estimated, medium-sized, open-economy New Keynesian model, cast it into a small-scale population vector autoregression, and assess whether best-practice structural identifications detect textbook “overshooting” after a monetary policy hike—i.e., an instant real appreciation that monotonically reverts. Our results include “delayed overshooting,” “exchange rate puzzles,” “forward discount puzzles,” and model-consistent overshooting. Identifications that regularly indicate open-economy anomalies in empirics likewise produce them in our controlled setup. Vice versa, identifications that prompt theory-conform conclusions in actual data do so in our experimental data. We infer that less empirical evidence may contradict canonical international macro theory than previously understood