Beyond the Zigzag Job Numbers
By Michael Englund
The U.S. employment data have followed a "feast or famine" pattern in the last several months. The latest example: the anemic 78,000 rise in nonfarm payrolls in the May employment report released June 3. That was well below the 274,000 surge (unrevised) reported in April and the 175,000 median estimate for May of economists surveyed by Action Economics.
As we had expected (see BW Online, 6/2/05, "Job Gains Fray in May"), the May data revealed a payback following the outsize gain in April, which may have been exacerbated by the five-week survey period for that month. Taking the longer view, however, the data still imply that the labor market continues to post healthy growth.
Among other highlights of the May report: The unemployment rate decreased to 5.1% (median estimate, 5.2%) from 5.2% a month earlier. The workweek held at a revised 33.8 hours (median, 33.8). Hourly earnings rose 0.2% (median, 0.2%). Manufacturing employment dipped 7,000, after a 9,000 decline in April. And the construction sector continued its hot streak, adding 20,000 jobs.
It's always risky to write off the weakness of a 78,000 payroll number as unrepresentative of the overall employment trend. But we think the latest report displays evidence that May's slack is little more than an extension of the oscillating monthly pattern in U.S. jobs gains since the middle of last year, where one month of strength has generally been followed by one of relative weakness. However, these swings are occurring around an otherwise unwavering trend, as the labor market continues to display firm growth, with the notable exception of the factory sector.
The pattern of strict monthly fluctuations in payroll growth extends way back to the weak jobs report of July, 2004, followed by a near-perfect pattern of swings around economists' median forecasts. We saw particularly big payroll gains of 282,000 in October, 300,000 in February, and 274,000 in April, and particularly small increases of 84,000 in July and 78,000 in May.
These swings provide little indication of the underlying quarterly growth trends. The picture becomes a bit clearer if we take a step back. Note that the average payroll gain of April and May was 175,000, and the average monthly payroll gain for 2005 is thus far 180,000. Those are both in the same ballpark as the 183,000 average monthly payroll increase in 2004.
An oscillating pattern is also evident in the household survey portion of the employment report, though the swings have been longer in duration, and the trend has actually gone in the opposite direction of late. The survey reveals a three-month string of hefty labor force and civilian employment increases over the March-May period -- with an average monthly jobs gain by this measure of 444,000. This followed a three-month string of weakness -- with an average monthly jobs decline of 50,000.
The poor payroll data for May translate to likely weakness in the month's remaining data reports, just as April strength boosted many of that month's figures. We now expect a below-trend 0.3% gain in personal income in May that follows the above-trend 0.7% in April. Both auto sales and retail sales have varied with the monthly jobs data. The estimated 0.5% drop in vehicle assemblies in May, which followed the even bigger 5.6% decline in April, likely helped to push the factory employment numbers lower, and retail employment rose by only 11,000 in May, following a 26,000 rise in April.
Average weekly hours worked in May rose a modest 0.1%, but this index is still poised for a remarkably strong 3.5% growth rate in the second quarter that would normally translate to a gross domestic product gain for the quarter that's well above our current 3.8% estimate. The fit is closer than it appeared to be after the robust April report, however, when we thought the second quarter's hours-worked growth might reach 4%.
In total, the employment data continue to provide a net positive signal for GDP growth in the second quarter, despite the disappointing May data.
A more persistently weak pattern is evident in the industrial production data, which should post a 0.1% May increase that follows the 0.2% drop in April, to leave a relatively anemic 1.1% growth rate for the second quarter. This has paralleled declines in most of the factory sentiment data and other factory reports.
Indeed, the outlook for factory employment appears bleak, just as the outlook for construction employment looks nearly bulletproof. These trends reinforce both themes in the market of a factory "soft patch" and a runaway housing sector. Growth in the latter has generally gained steam since the beginning of 2003, and activity in the housing market through the seasonally important second quarter shows no sign of a slowdown.
Overall, while the nonfarm payrolls and household survey data look like they're zigging and zagging, the movements appear to be above and below a stable trend. The key to the Federal Reserve policy outlook beyond the expected quarter-point tightening in June will be this month's employment report, where we at Action Economics now expect an innocuous 175,000 gain in nonfarm payrolls.
Nevertheless, we still believe the labor market and the economy are stronger than the May headline employment figure suggests, and we suspect the Federal Open Market Committee won't pause with its series of rate hikes until the Fed funds target rate reaches 4%.
Englund is chief economist of Action Economics