Missing Japan data after 2000
The análisis looks solid.
What's the confidence interval on this estimate? The margins seem awfully tight.
Is the anomaly as a share of total here?