Michael Lucas comments on July 1st, painting a grim picture of Wisconsin employment. For comparison, here are some other aggregate numbers, including the Philadelphia Fed’s coincident index, released today.

Figure 1: Wisconsin Nonfarm Payroll Employment (dark blue), Philadelphia Fed early benchmark measure of NFP (pink), Civilian Employment (tan), real wages and salaries deflated by national chained CPI (sky blue), GDP (red), coincident index (green), all in logs 2021M11=0. Source: BLS, BEA, Philadelphia Fed [1], [2], and author’s calculations.

The coincident index shows rapid growth in recent months (revising up May’s number). This is interesting to the extent the coincident index relies mostly on labor market indicators.

Mr. Lucas’s point in his article is that more reliable indicators of employment have shown decreases in jobs, viz:

The loss of jobs over the past year can be more easily seen by referring to the BLS’s QCES survey—a lagging indicator of the economy that is among the most accurate and comprehensive surveys of the marketplace. According to data from that survey, Wisconsin gained a net total of 8,099 jobs from 2023Q1-Q3 (Q4 data is not available yet). However, in the second and third quarters, Wisconsin lost a net total of 20,207 jobs.

He presents this table in support of this argument.

Source: Lucas (2024).

Apparently, Mr. Lucas is not aware that the QCEW data is not seasonally adjusted, so that really only the q3/q3 calculation makes much sense.

Personally, I’m not familiar with the QCES (anybody who is, please tell me what it is). I am familiar with the QCEW, which is not a survey, but a census (hence the “C” in QCEW).

I plot the changes in employment from BED (a survey), QCEW and from the CES.

Figure 2: Change from 2021Q4 in Wisconsin private nonfarm payroll employment from Current Population Survey (CPS) (blue), in Quarterly Census of Employment and Wages (QCEW) (green), and from Business Employment Dynamics (BED) survey (tan), all in 000’s. QCEW series seasonally adjusted by author using X-13 (log transform) on monthly data.  Source: BLS, and author’s calculations.

To highlight the importance of seasonal adjustment, compare nsa QCEW vs. sa QCEW…

Figure 3: Wisconsin private nonfarm payroll employment from Current Population Survey (CPS) (blue), from Quarterly Census of Employment and Wages, not seasonally adjusted (QCEW) (light green), from Quarterly Census of Employment and Wages, (QCEW) (green), all in 000’s. QCEW series seasonally adjusted by author using X-13 (log transform) on monthly data.  Source: BLS, and author’s calculations.