U.S.: The Job Market Is Stronger Than It Looks
November's employment report, showing a meager 57,000 increase in U.S. payrolls, was a real stinker. That was still true even after accounting for the strike on the West Coast of some 23,000 grocery-store workers. However, given all the other data showing the improving tone of the labor market, the monthly payroll numbers may not be giving an accurate picture of job growth. That's especially true when one considers that the economy is likely to post growth at an annual rate of about 5% for 2003's final three quarters.
To be sure, productivity growth has been soaring -- jumping at an annual rate of 9.4% in the third quarter alone -- and if businesses are increasing output through better efficiencies, they don't have to add to payrolls. But productivity this quarter is already showing signs of slowing, so more workers are needed to meet the faster pace of demand accompanying the economy's new vigor.
Even the Federal Reserve is more optimistic. After announcing it had kept interest rates unchanged, the Fed said: "Output is expanding briskly, and the labor market appears to be improving modestly." That assessment was more upbeat than its August or October views.
Policymakers also said that the risks of deflation have diminished. The bond market read that as a sign short rates might rise sooner than expected, despite the Fed's continued belief that policy can remain accommodative for "a considerable period." Both bond traders and the Fed will be keeping an even closer eye on the job markets for signs of potential price pressures as 2004 progresses.
THAT'S WHY CARVING OUT an accurate view of the labor markets is so crucial. But November's job growth is at odds with several other job-market indicators. For example, weekly unemployment claims are down to levels consistent with much larger job gains. Employment indexes from the nation's purchasing managers, which measure the breadth of hiring, are at levels not seen since before the recession. And a survey by the National Federation of Independent Business says hiring plans among small businesses in November were comparable to the strong readings of the 1980s expansion.
Perhaps most important, the Labor Dept.'s other survey of employment, taken by canvassing households instead of businesses as with the payroll data, shows sharply greater job growth in 2003 than the more widely followed survey of company payrolls. So far this year, the household survey, on which Labor's calculation of the unemployment rate is based, shows employment has grown by 2.2 million workers. The payroll survey says jobs are down by 24,000 (chart).
One difference in the coverage is that the household survey counts people who say they are self-employed, while the payroll data do not. Contrary to popular notions, however, this is not the major difference in the two accountings. True, the count of self-employed workers is up substantially -- by more than 330,000 this year, and by nearly 700,000 since the recovery began two years ago. But even excluding the self-employed, the household survey shows 2003 job growth of more than 1.8 million, possibly because that survey better captures the hiring at small companies, which sometimes falls through the cracks of the business survey.
In response, Labor has instituted improvements in its payroll-survey methods, including new sampling techniques to collect more accurate data on business startups and the hiring they generate, a key source of undercounting in past recoveries. Plus, Labor officials say that after all revisions are in, the two surveys tend to match up more closely.
Revisions are coming soon. In January, Labor will release its annual retooling of the household data and follow that in February with benchmark revisions to the payroll data. The guessing is that the revised job growth will ultimately settle somewhere in the middle of the two surveys, with the recent payroll data looking stronger.
IT WOULDN'T BE THE FIRST TIME that revisions gave the job markets more muscle. The initial payroll data reported by Labor has a history of underestimating job growth early in a recovery when hiring begins to take off. For example, in the second year of the recovery from the 1990-91 recession, payroll gains were first reported to be only about 300,000. But a series of revisions now shows job growth was a much higher 1.6 million. In the third year, job gains were originally reported as 2.2 million, but were later revised up to 3.2 million.
Regardless of what next year's revisions show, job growth seems bound to strengthen in coming months, given the recent trends in several forward-looking indicators. First, based on past correlations, the four-week average of new unemployment claims, at 363,000 at the end of November, is consistent with monthly job growth of around 150,000.
Second, temp hiring tends to lead hiring of full-timers, and temp employment since May has jumped by 166,000, the best seven-month rise since early 2000. Third, the workweek last month increased to 33.9 hours, the longest in more than a year, and in manufacturing the workweek rose to 40.8 hours, the longest in three years. As demand picks up, companies tend to stretch out their workweeks, which is typically a precursor to new hiring.
Finally, productivity is following its normal business-cycle behavior. Its recent surge is typical in the early stages of a recovery, as output picks up in advance of new hiring. Now, with businesses more confident about the recovery, hiring is picking up, and productivity will slow.
IT'S ALREADY HAPPENING. Fourth-quarter output, based on gross domestic product, is expected to come in at about a 4% annual rate, but data through November suggest that hours worked are rising at an annual rate of about 3%. That would be the first increase in hours worked in three years, and the largest rise in four years (chart). As a result, fourth-quarter productivity is making little headway, and a slower pace will continue in 2004.
Another reason why job growth will pick up is the rebound now under way in manufacturing, which will help to stanch the enormous loss of factory jobs. Since the end of last year, manufacturing has let go of 478,000 workers. The service sector, which employs 83% of all payrolls, has added 336,000.
Factory layoffs are already moderating. In the past three months, declines have averaged 17,000 per month, down from a 55,000 pace in the first half of the year. And next year, manufacturers expect to lift their hiring, at least modestly, says the Institute for Supply Management in its yearend survey. More broadly, the ISM says, "manufacturing purchasing and supply executives are optimistic about their organization's prospects for the first half, and they predict additional growth during the second half of 2004."
Unemployed factory workers have been waiting for three years for the payroll ranks to start rising again. Now, with demand set to keep expanding into 2004, the outlook for factory jobs -- as well as service employment -- is improving, no matter what the November job report seemed to imply.
By James C. Cooper & Kathleen Madigan