Ed Dolan's Econ Blog

US Unemployment Edges Down to 7.3 Percent, a New Low for Recovery

The U.S. unemployment rate edged down from 7.39 percent in July to 7.28 percent in August, according to data released today by the Bureau of Labor Statistics. The decrease did not, however, reflect an across-the-board strengthening of the labor market. According to the BLS household survey, the civilian labor force, the number of unemployed, and the number of employed all decreased slightly for the month, both before and after seasonal adjustment. The labor force participation rate and the employment-population ratio also decreased on the month.

The BLS also publishes data on a broader measure of unemployment and undermployment known as U-6. That measure takes into account people who are working part-time but would prefer full-time work, and so-called marginally attached workers. The latter incude people who have not looked for work because they think none is available and people who would like a job and are available for work, but who did not look for work in the previous four weeks because of study, family responsibilities, or other reasons. Both involuntary part-time workers and marginally attached workers decreased for the month, bringing U-6 to 13.7 percent. As the next chart shows, that also was a new low for the recovery.

According to the separate survey of business establishments, the number of payroll jobs grew by 169,000 during August. The establishment survey excludes farm workers and the self-employed, does not correct for workers holding two jobs, and differs in other details of methodology. It is not unusual for the household employment data and the payroll jobs data to point in opposite directions. Previously reported payroll jobs gains for June and July were revised downward by a total of 74,000 for the two months. As the next chart shows, the August job growth was slightly below the average of 184,000 per month of the preceding year.

Most of the new payroll jobs were in the service sector, with retail trade, professional and business services, and healthcare showing some of the strongest gains. Manufacturing showed a small increase in jobs while construction was flat. Government employment gained, largely on growth of local government education. Federal government employment was unchanged.

Not too much can be read into any one month’s jobs numbers. The surveys on which they are based are subject to sampling error and later revisions. Considering these caveats, the August employment situation is consistent with pattern of the past two years—slow but steady growth with a considerable distance to go before the economy reaches anything that we could reasonably describe as “full employment.”


8 Responses to “US Unemployment Edges Down to 7.3 Percent, a New Low for Recovery”

SchofieldSeptember 6th, 2013 at 1:56 pm

Surely these statistics ought to be issued by a renamed department the United States Department of Economic Lunacy not Labor! Human resources left to rot by a Congress that is utterly clueless how to run a successful economy!

JanaSeptember 8th, 2013 at 3:46 pm

Interesting statistics – here in Germany the situation looks right now a bit better. Hope the same will hapen for the US soon.

windrivenSeptember 9th, 2013 at 9:48 am

The decline in U6 is encouraging but the slight pullback in the total number of employed is a little disturbing. As Dr. Dolan points out GDP, at least nominal GDP, is growing modestly. But it is a little unclear from whence the demand that is driving that growth. Businesses do not invest in labor and equipment without demand for their products and services. Government efforts to stimulate demand in the US have been rather less than wildly successful. Absent a growing appetite for US products abroad it is hard to see robust recovery in the near future.

Ed Dolan EdDolanSeptember 9th, 2013 at 10:18 am

You are right, the fiscal drag from the steadily decreasing government contribution to GDP (government consumption and gross investment) is undermining demand and employment. According to the latest revision of the GDP data, US exports were surprisingly strong in Q2, but with so many trading partners slowing down, that could change.

rjsigmundSeptember 9th, 2013 at 12:11 pm

ed, i had a problem with this report…there is a discrepancy of almost a million jobs between the unadjusted employment change data from the two BLS surveys, as the seasonal adjustments subtracted more than 200,000 jobs from the establishment survey and added nearly 500,000 to the household survey…

rjsigmundSeptember 9th, 2013 at 12:11 pm

the unadjusted data from the establishment survey shows payrolls jobs increased by 378,000 from 135,583,000 in July to 135,961,000 in August; after which the BLS seasonal adjustment increased the payroll jobs to 135,964,000 in July and 136,133,000 in August which thereby reduced the jobs change from July to 169,000; meanwhile, the raw, not seasonally adjusted household data shows employment dropped 604,000, from 145,113,000 in July to 144,509,000 in August…however, the action of the seasonal adjustment on the household survey was in the opposite direction, in that it reduced the negative change of employment loss to 115,000, changing the monthly change from 144,285,000 in July to 144,170,000 in August….thus the seasonal adjustment subtracted 209,000 from job gains in the establishment survey, but it added to 489.000 those counted as employed in the household survey…we know the establishment survey is more accurate, but there is no reason for the seasonal adjustments on employment to move in opposite directions by that magnitude in the same month…

Ed Dolan EdDolanSeptember 10th, 2013 at 9:33 am

We'd have to look back over the years to see if this is an exceptional discrepancy. However, although the surveys roughly track over the long run, there is no particular reason for them to line up on a month by month basis. The methodologies are quite different. The establishment survey is based on actual data and adjusted when new data come in, whereas the household data is a sample survey subject to sampling error but revised only annually, when the data are rebenchmarked. Also, the household survey includes farm workers and self employed, who may have very different seasonal characteristics. Finally, the establishment survey double-counts multiple jobs whereas the household survey doesn't. I wouldn't make too much of it unless a discrepancy this large were to persist month after month.