Time to correct a little silliness on what happened to women's employment in April. Several reports claimed that men got most of the jobs in April. This is based on the household survey.

The month to month changes in this survey are highly erratic and largely driven by measurement error.

To take a couple of examples, in February of 2014, women's employment was reported as falling by 269k. Can anyone think of an event that would cause this sort of plunge, I can't? In any case, in March it rose by 677k.

In October of 2017 women's employment reportedly fell by 322k. That doesn't sound as bad when we consider that it reportedly rose by 490k in September.

The takeaway is that the month to month movements in the household survey are highly unreliable. If we look over a longer period, we can be pretty comfortable it is giving us a story, but single month movements need a large grain of salt.

We do have another data source on women's employment. The Bureau of Labor Statistics survey of employers reports the number of women on employer's payrolls each month.

This survey showed that 161k of the 266k jobs created, or just over 60 percent, went to women. This survey has error also, so I wouldn't swear that 60 percent, or even more than 50 percent, of the new jobs went to women.

But this figure is hugely more accurate than the changes reported in the household survey. If we want monthly changes, this is the better survey to rely upon.