Employment situation: March 2025
The first ghost post (ghoast poast) — and it's not that interesting. Which is good! Unemployment looks like what we expected last month (and the month before that, and the month before that ...). The last five data points (the last ones that fall in the gray band in the graph at the top) show a log rate of decline of
−0.18 ± 0.14 y⁻¹
which, while fairly uncertain, encompasses the expected rate −0.1 y⁻¹. If the current administration doesn't cause a recession (i.e. trigger the Sahm rule again) and this log rate continues to converge on −0.1 y⁻¹ that would be a pretty good feather in the DIEM's hat. I'd add a footnote at this point if I was still on substack, but that's not a feature of ghost — so here it is inline: the log rate of −0.1 y⁻¹ is a prediction of the DIEM, but rejoining the "post-pandemic no-stimulus counterfactual" hypothesis is not and points to restorative forces.
Another hypothesis that is going strong is the noise floor — and we're still at it:
All in all, we're not seeing any particular deviations yet from the expected paths. Since it's possible this administration might start tampering with the economic data, the DIEM might be a good way to detect malfeasance.
