This isn't a very good employment report. Here are three views of the report, starting with Dean Baker, plus links to more discussion at the end:
Economy Sheds Jobs, Unemployment Stable, by Dean Baker, CEPR: "For the first time since data has been kept, manufacturing is less than 10 percent of employment."
The establishment survey showed the economy losing 17,000 jobs in January, the first reported job decline since August of 2003. Over the last three months, the economy has added a total of 125,000 jobs, with the private sector adding just 99,000 jobs. The household survey showed little change in the unemployment rate, although it rounded down to 4.9 percent for January compared to 5.0 percent in December.
The job loss in the establishment survey was driven by losses of 28,000 jobs in manufacturing, 27,000 jobs in construction and 26,000 state government education jobs. The latter is likely due to a faulty seasonal adjustment and will probably be reversed in future months, but the declining employment in construction and manufacturing is very real.
Residential construction has lost 189,300 jobs since July, 5.8 percent of employment in the sector. Employment in the non-residential sector has also been drifting downward in the last three months, indicating that the boom in this sector is over. ...
There were few sectors showing notable job gains in January. The two exceptions were health care and restaurants. The health care sector added 27,100 jobs, while the restaurant sector added 14,800 jobs. Over the last three months, these two sectors together have added 130,100 jobs, an amount equal to 131.4 percent of the job growth in the private sector as a whole.
The data in the household survey largely reinforce the picture of a weak labor market. There was little change in unemployment rates for most demographic groups, although the January data indicate that a reported jump in December in the unemployment rate for black teens was not a fluke. The January data show a black teen unemployment rate of 35.7 percent. This is up from a 27.9 percent rate reported for October.
The number of people involuntarily working part-time rose again, hitting the highest level since October of 2004. The average and median duration of employment spells both increased, as did the percentage of the long-term unemployed. The number of discouraged workers also showed a year over year increase, continuing the recent pattern.
By age, employment growth continues to be heavily skewed toward older workers, with people over age 65 showing an employment gain of 287,000 jobs, while employment dropped by 250,000 for other age groups. ...
This employment report should be sufficient to remove any doubt that the economy is in very serious trouble and most likely has already entered a downturn. In addition to the loss of jobs, there was also a reported decline of 0.1 hours in the average workweek, leading to a decline of 0.3 in the index of aggregate weekly hours. The decline in hours worked, which showed up most clearly in non-durable manufacturing, suggest that more layoffs are on the way. In addition, wage growth has slowed to crawl, averaging just 2.3 percent over the last quarter. This is well below the rate of inflation.
It is also important to remember that the birth/death imputation is likely overstating the number of jobs created in new firms. The revision for last year lowered total employment by 376,000 or 31,000 per month. It is likely that the current data will be revised downward by a comparable amount. In short, the picture is probably even worse than the data now show.
The insignificance of zero, by Paul Krugman: So the new labor report is out, and it says that nonfarm payrolls actually fell last month. On the other hand, employment growth for December was revised up. You shouldn’t take any of this seriously. For one thing, seasonality is a big problem. There’s normally an employment bulge in December, as stores and others bulk up for the holiday, then a slump in January as they let the extra workers go. The BLS tries to adjust for these seasonal patterns, but because the pattern is always changing, it’s an imperfect process. A better guide is probably to average the last 2 or 3 months. What you get then is that employment is still growing, but v-e-r-y s-l-o-w-l-y. In particular, employment growth is well short of what’s necessary to keep up with population growth. So even though it’s premature to say that jobs are shrinking, as a practical matter this makes no difference: the truth is that the jobs picture looks moderately dire.
The WSJ Economics Blog:
One Weird Employment Report , by Greg Ip and Kelly Evans: Today’s jobs data were a melting pot of fascinating and conflicting signals on the labor market.
Of course, the 17,000 drop in nonfarm payroll employment is the most eye-catching aspect. The weakness in the payroll survey was corroborated by a decline in the average work week, to 33.7 hours from 33.8 in December, and a slowing in hourly wage gains, to 0.2% from 0.4%.
But the separate household survey showed a whopping increase in employment of 635,000, when new updated population controls are applied to both the December and January data. ...
In December, just the opposite occurred: nonfarm payrolls rose a respectable 82,000, upwardly revised from the initial estimate of just 18,000. But that month, household employment plunged 436,000.
The household survey’s signal of strength was ratified by the unemployment rate, which dropped to 4.93% from 4.98%. The employment to population ratio jumped to 62.9% from 62.7%. These ratios are more reliable than the absolute numbers for employment and unemployment which are whipped around by the small sample size; they suggests a marginal improvement in the labor force.
It would be easier to draw conclusions if other data were pointing universally in one direction or another. But they’re not. Unemployment claims pointed to strength during January, until the last week, when they pointed to sudden weakness. The ADP payroll report pointed to remarkable strength. But regional purchasing manager surveys pointed to weakness. (We’ll get a broader look with the national ISM index due out at 10 a.m.)
Meanwhile, revisions to the payroll survey continue to make the report difficult to read. The Labor Department revised up their estimate of December job creation by 64,000 jobs to 82,000 and revised down their November estimate by almost half, to just 60,000 jobs created that month. And there were a whole slew of so-called “benchmark” revisions showing the total seasonally-adjusted level of nonfarm employment in December 2007 was 376,000 lower than first thought...
Plus, this isn’t the first time the government has reported a negative monthly number. Their initial estimate of August job creation, released in early September, saw overall employment drop by 4,000 jobs. Coming as it did on the heels of the August credit crunch, the Wall Street Journal reported that “The jobs report stoked fears of recession, sending shockwaves through financial markets.”
That number was subsequently revised to show a gain of 89,000 for the month of August.
Fast-forward to today, and it’s clear the labor market is softening. But it’s also clear the monthly jobs data should be taken with a grain of salt.