Some key labor market figures simply aren’t adding up these days, leaving Federal Reserve officials in doubt about what’s happening in critical parts of the economy.
In some cases, the participation of illegal immigrants in the work force may be roiling the data. With the economy operating at or close to full employment, so-called unauthorized workers have become an important source of new hires for many businesses.
Foreign-born workers – an unknown proportion of whom were in the country illegally – accounted for 45 percent of all employment growth last year, according to the Bureau of Labor Statistics.
The question is how unauthorized workers and their employers respond to the household and business establishment surveys on which the highly important employment, unemployment and payroll job figures are based. It may be that those responses have been changing over time.
Housing-industry employment is at the top of the list of the data discrepancies.
Housing starts have plummeted, and the number of units under construction has fallen 15 percent in the year ended in April. However, the BLS’ monthly survey of business establishments showed 3.3 million people still at work on residences last month, only a 4 percent drop from a year ago.
Economists at the BLS and the Fed know that shifts in the economy often affect the employment data with a lag. This discrepancy, though, seems more than just a matter of a lag.
According to Ray Stone of Stone & McCarthy Research Associates, one possible explanation is that thousands of unauthorized workers employed by very small subcontractors have lost their jobs. In some cases, the workers may have been paid “off the books,” so their departure may not be reported. And it may be that such businesses are underrepresented in the sample of companies reporting payrolls each month, Stone said.
If so, then payroll employment may have been higher than it appeared to be before the housing slump began.
Another puzzling piece of payroll data emerged earlier this year, when the BLS released its annual benchmark adjustment for March 2006. The adjustment, which is based on quarterly federal unemployment insurance payments by employers, was an unusually large one, adding 752,000 more workers to payrolls as of that month.
A significant portion of the adjustment came in three states, California, Texas and Louisiana, with the second-largest impact in construction. The first two states have a high proportion of foreign-born workers, many of whom are employed in construction.
Stone published an analysis last year suggesting that unauthorized workers who may not have been reported on payrolls earlier were put there as a result of a “slew of actual and prospective legislation that increased penalties on employers that knowingly hired illegal immigrants.” That is, the prospect of penalties caused employers to seek documentation for workers, some of whom probably had fake papers.
Now, however, there are new questions about the accuracy of the payroll data.
First of all, the BLS routinely adds a substantial number of workers to its monthly payroll figures to make up for the fact that in a dynamic economy, as new companies start up and old ones close their doors, the sample of businesses providing payroll data routinely undercounts the number of workers.
This “birth/death” model doesn’t take into account cyclical fluctuations in the economy, and on May 18, Stone told his clients that as a result, payrolls may have been growing more slowly than reported since the benchmark adjustment.
Stone noted that on May 16, the BLS released its Business Employment Dynamics – or BED – report for the third quarter of 2006. The report provides data for changes in payroll employment using the unemployment insurance filings from businesses and sorts it according to whether the businesses are new or expanding, or contracting or closing. In other words, it tracks the information the “birth/death” model is supposed to provide in a more timely fashion.
“The new data revealed a seasonally adjusted third-quarter private payroll gain of only 19,000, in sharp contrast to the BLS’ published monthly payroll increase of 498,000 for the quarter,” Stone said.
For the 2006 second and third quarters together, the BED data show a payroll increase of 485,000, compared with 808,000 for the regular payroll figures. For the six months, that’s the difference between monthly payroll gains of 80,000 versus 135,000, Stone said.
Many economists and more than a few Fed officials have been puzzled why various labor market measures seem to have been so little affected by the downshift in economic growth to about a 2-percent annual rate or less in the middle of last year. The BED report hints that maybe the payroll data at least will show just such an impact in the next benchmark revision.
And then there is the 4.5-percent unemployment rate. Why has it remained so low?
Well, suppose many unauthorized workers – those are the illegal immigrants – don’t show up in the household survey on which those monthly figures are based. Yet when they have a job they are, in reality, part of the work force. And when they lose a job, they don’t get counted as unemployed. They just aren’t part of the work force anymore.
There’s another piece of evidence that that has been happening. Remittances from Mexican workers in the U.S. to their families at home have dropped significantly in recent months.
John M. Berry is a Bloomberg News columnist.
