Word for the day: Rotation group bias. = Labor market survey responses differ when folks are re-surveyed. It matters because the government shutdown means there are more inexperienced respondents in November's survey. Will likely push the *measured* unemployment up a tad.

Also: Typically only one-eighth of the sample is new, allowing the BLS to use panel data methods (effectively fixed effects) to account for sampling error. But this time a quarter of the sample is new, so this month's estimates will also have greater sampling error.

These are just two examples of ways in which the government shutdown will have echo effects that make it difficult to parse the official data. We're out of the statistical blackout, but still in a deep fog.