We should fix this in PolicyEngine-US, but essentially: sometimes we get -1 decile values for a few individual records. Probably because of negative income. But then when all records in one decile have zero weight, we get an error when taking the mean of something grouped by decile.
Fixes #1545
We should fix this in PolicyEngine-US, but essentially: sometimes we get -1 decile values for a few individual records. Probably because of negative income. But then when all records in one decile have zero weight, we get an error when taking the mean of something grouped by decile.