Two Clocks, One Verdict

Employment reached 159,075 from 158,913, a monthly increase of 162. Public jobs stand at 23,323 against 23,535 a year earlier, a change of -212. These levels and changes are in thousands. The signs disagree, but so do the clocks. A monthly gain and an annual contraction are not rival answers to the same question. Don can take the employment increase as confirmation that expansion is carrying the labor market with it.

His position spends the distinction between a total and its composition. The aggregate supports the narrower statement that payrolls increased over the month. It does not establish whether public employment rose or fell during that same interval. The annual public jobs comparison cannot supply the missing monthly answer, however neatly its minus sign fits the argument. The opposite conclusion has the same problem.

Public employment below its year-earlier level does not establish that the latest monthly increase is concealing a fresh contraction there. That would require a comparison the quoted figures do not make. An annual decline can coexist with a recent increase, just as a monthly gain can coexist with a decline over a longer interval. Neither possibility is established here. Both employment figures are provisional, so revisions can also change the apparent shape of the disagreement.

The next monthly payrolls release is the next test. The first requirement is to compare total and public employment over matching intervals, then examine the revisions. If both move together on that basis, the supposed split loses its evidence. If their directions still differ, Don must concede that the aggregate is an incomplete description of the labor market. Even then, the contrast would describe its composition, not explain what caused it.