The jobs report for October due out Friday may be bleak. It might even be scary. The unemployment rate could jump by the most in three years. Hiring may slow from an already weak pace.
The ugly figures will reflect the government's partial shutdown, which coincided with 16 days in October. The trends for the job market will likely reverse themselves in coming months.
"It's going to be a very messy report, and I don't think we think should take it at face value," said Jennifer Lee, an economist at BMO Capital Markets.
Economists warn that the unemployment rate could surge as high as 7.5 percent from 7.2 percent in September. That would be the steepest one-month rise since 2010.
The number of jobs added in October could slow to roughly 120,000 from the 148,000 added in September. That isn't healthy. In the first nine months of this year, the average job gain was 180,000.
The shutdown will be mostly to blame. But its effect on the data won't be easy to tease out. Economists have all but thrown up their hands trying to forecast Friday's figures or to suggest what they might mean. However the numbers turn out, the distortions mean the monthly jobs data will be less useful in gauging the economy's health than they normally are.
"We have much less confidence in these numbers than usual," economists at Bank of America Merrill Lynch wrote in a note for clients.
Why the confusion?
Consider how the jobs report is compiled: It's derived from two separate surveys. Each survey will be affected differently by the shutdown.
One is a household survey. Government workers ask adults in a household whether they have a job. Those who don't but are looking for one are counted as unemployed. That's how the unemployment rate is calculated.
The other is a payroll survey. The government asks mostly large companies and government agencies how many of their employees worked or received pay, typically during the second week of the month. This survey produces the number of jobs gained or lost.
Suppose you're a federal worker who was furloughed by the shutdown. The payroll survey would consider you employed. But the household survey would count you as unemployed.
Why the disparity?
Because furloughed federal employees received back pay for the time they didn't work. So for the purposes of the payroll survey, they were employed. The same is likely true for government contractors who were temporarily laid off. Many were probably paid for at least part of the time covered by the payroll survey. So the payroll survey will consider them employed.
That's why October's job gain isn't expected to drop much.
The household survey takes a different approach: It will count both the federal workers and the contractors as unemployed because they weren't working during the survey period.
The shutdown furloughed about 450,000 federal employees in the second week of October. If the number of unemployed rises by that much in October's jobs report, the unemployment rate could reach 7.5 percent.
There will be other distortions. Some people with private employers might have been affected. Examples would be employees of hotels and restaurants near national parks that were closed. Those workers might have been temporarily laid off, thereby boosting the unemployment rate.
Yet as long as they were paid for even one day of work during the survey period, the payroll survey would count them as employed.
"It is basically impossible to determine how many private-sector workers were affected by the shutdown," economists at JPMorgan Chase wrote in a note to clients.
The shutdown likely stopped the government from hiring for roughly two weeks. Goldman Sachs estimates that this fact lowered government employment by 12,000 in October. With the shutdown over, those jobs will be added over time.
In addition, some workers, in both the government and private sector, might have had their hours cut during the shutdown. Their reduced hours might have boosted the number of part-time workers in October.
The shutdown's effects will be largely unwound by November. But then the bounceback from October could distort November's jobs data.
A clear picture of the job market's health may not emerge until January, when December's jobs report will be released.