Like Diogenes’ quest for an honest man, seeking truth—or even consistency—in the monthly employment situation report may prove an endless task.
Friday’s report on the August employment situation at the same time confounded analysts twice with a less than robust 169,000 month-over-month increase in payrolls and a staggering 58,000 reduction in the initial report on payrolls for July.
For those who keep track, June payrolls were initially reported to have increased 195,000 but with Friday’s report have now been revised down twice: first to 188,000 and Friday to 172,000. Who knows how low the July (or the August) payroll numbers could go.
Payroll report revisions are, to be sure, nothing new. The Bureau of Labor Statistics (BLS) initially reported payrolls for March had increased by 88,000 in the release issued at the beginning of April. One month, later March payrolls were reported to have increased 138,000, and one month after that 142,000.
That said, the 58,000 downward revision is unusual. The initial payroll report hasn’t been reduced by that much since June 2010, when the initial payroll estimate of a loss of 125,000 was changed to a drop of 221,000 jobs and then, one month later, to a loss of 175,000 jobs—suggesting it’s no easier to count going down than it is going up.
BLS said 74.1 percent of payroll surveys were returned in time for the Employment Situation release with the first set of July numbers. The collection rate was lower for January (72.4 percent) and for August (also 72.4 percent). January’s report of 157,000 new payroll jobs was reduced to 119,000 in the first revision before settling at 148,000 in the final report.
The “final” report may not be final. BLS said it would release its annual “benchmark” revisions at the end of September, which will change the numbers yet again.
While it may not be easy to reconcile the numbers in the employment situation report with themselves, it is even more difficult to reconcile them with other economic indicators.
For example, builder confidence, according to the Housing Market Index of the National Association of Home Builders, has increased 34 percent in the last three months, vaulting into positive (above 50 on a scale of 1 to 100) territory in June for the first time since April 2006. In those same three months, residential construction employment is down 0.2 percent, a loss of about 1,100 jobs.
And, even though data on residential permits and starts show gains only in multifamily activity, jobs in the multifamily contractor sector are also down.
BLS numbers don’t seem to square with recent surges in sales of new and existing single-family homes. While new home sales are up on average better than 20 percent year-over-year and existing-home sales up about 12 percent, the number of mortgage brokers has increased just about 10 percent and appraisers just about 6 percent. The number of underwriters is up just 1.2 percent.
If those numbers aren’t enough to raise questions about the depth of the housing recovery, consider that in the last three months, payrolls have climbed an aggregate 445,000, while at the same time, the number of multiple jobholders has increased 147,000, which would mean the new payroll jobs themselves didn’t do as much to reduce unemployment.
Of the new jobs, about 221,000, almost 50 percent, were in the two lowest paying industry sectors—leisure and hospitality (83,000) and retail (138,000) with average weekly earnings of $352 and $523, respectively—numbers and earnings far from enough to support a housing recovery.
The August employment situation report wasn’t the first set of data to suggest trouble on the horizon for the housing recovery. The Case-Shiller home price index for June—the most recent—showed continuing, albeit slower, house price gains, pushing affordable homeownership still further from low paid workers.
That is, until the numbers change again.
By Mark Lieberman