As one who has often reported the monthly payroll jobs breakdown, I am struck by the fact that these categories are the ones that have accounted for job growth for year after year. How can this be? How can Americans, who have had no growth in their real incomes and who are foreclosed from their homes and maxed out on credit card debt, car payments, and student loans, spend more every month in bars and restaurants? How can a few service areas of the economy grow when nothing else is?
The answer is that there were not 192,000 new jobs. Statistician John Williams estimates the reported gain was overstated by about 230,000 jobs. In other words, about 38,000 jobs were lost in February.
There are various reasons that job gains are overstated and losses understated. One is the BLS’s “birth-death model.” This is a way of estimating the net of non-reported new jobs from business start-ups and job losses from business shut-downs. During recessions this model doesn’t work, because the model is based on good times when new jobs always exceed lost jobs. On the “death” side, if a company goes out of business because of recession and, therefore, doesn’t report its payroll, the BLS assumes the previously reported employees are still in place. On the “birth” side, the BLS adds 30,000 jobs to the monthly numbers as an estimate of new start-ups.
Williams estimates the “death” side is really reducing employment by about 200,000 per month, and the “birth” side is stillborn. Therefore, “the BLS continues regularly to overestimate monthly growth in payroll employment by roughly 230,000 jobs.”