The payroll report disappointed again, though it was deficient in ways other than are commonly described. The monthly change is never a solid indication, good or bad, as the BLS’ statistical processes can only get it down to a 90% confidence interval, and a wide one at that. It means that any particular month by itself specifies very little, except under certain circumstances. This month just happens to be one, pointing in the direction of more useful interpretation.

At just +98k, the headline change in the Establishment Survey was about half of what was expected. In contrast to last month, such a low delta is meaningful in only one respect – given the confidence interval the BLS cannot say with significance that the payroll report for March 2017 was even positive. It still may have been lackluster or even a good one, but it also may have been negative.

It is therefore consistent with the rest of our economic findings. Though there is some improvement in certain areas, it isn’t meaningful so as to indicate what everyone seems to think it does. Instead, the US economy remains weak even if this year’s weakness isn’t as immediately dangerous as last year’s. The payroll reports fall well within that range, especially when viewed beyond a single month.

Because of that 98k as well as a -38k revision to last month (February 2017), the 6-month average of estimated payroll gains is now just 163k. That is 102k per month less on average than the estimated peak in the middle of 2014, a serious indicated slowdown. It is also the lowest average gain in four and half years, going all the way back to late 2012 and the initial slowdown.

Though there are all kinds of discontinuities (population figures) and irregularities (SNAP) with the labor force estimates, cutting through them there does not seem to be much enthusiasm for entering the labor market (apart from regulatory changes to SNAP). The lack of acceleration of new entrants is consistent with both the currently weak state as well as its chronically weak post-crisis baseline.