In running the 100-yard dash or weighing diamonds, fractions are important. In measuring national employment, a change in the tens of thousands sometimes means nothing.
Table A of the Bureau of Labor Statistics’ (BLS) Oct. 7 employment press release showed total nonfarm payroll jobs in September down 35,000 from August, reflecting the effect of Hurricane Katrina. (Economists had predicted a drop of up to 200,000.) Headlines around the country labeled it a nationwide decline in jobs, the first in more than two years.
But there was no August-September decline in the officially reported job count. Data collected from samples have sampling error, a band within which plus or minus changes are not different from zero. The Bureau’s analytical Table 1, published along with the jobs data, shows a total job change for September had to be at least 103,000 to be statistically significant. Anything smaller is not significant and doesn’t count. Thus, correctly assessed, the job situation was unchanged. There was no net job loss. The job market didn’t decline; it was stagnant. Heads up, media.
The same BLS press release showed total civilian employment, based on the government’s household survey, was 17,000 lower in September than in August. But again, the change was within sampling error and so was not statistically significant.
To capture as much of the employment picture as possible in Kartina-affected areas, the BLS temporarily modified its usual estimating procedures. While the modifications doubtless improved the September count’s accuracy, sources of bias remained. For example, in the payroll count, nonresponding firms were assumed to have zero employment. Normally, nonresponders would be assigned the same monthly change as others in the sample, so this modification risked overstating the estimated loss in jobs.
For household data, units in the hurricane area that were uninhabitable or whose condition was unknown were removed from the sample and households that could be successfully interviewed were given higher weights. To the extent the people interviewed were likelier to be employed than those missed, the measured employment loss would be understated.
While the reported effect of Katrina on September employment was less than expected, Gulf area job loss was still substantial, though it’s difficult to separate the hurricane’s contribution from changes in the overall economy. (The payroll and household surveys’ reference periods were prior to Hurricane Rita, so Rita didn’t affect the September data.)
The job market was on a sharply rising trend prior to Katrina. From January to August, payroll jobs rose an average of 215,000 monthly. And in September the economy and the labor market outside the Gulf Coast were doing well. Assuming the pre-Katrina trend in job growth would have continued into September had there been no hurricane, payroll jobs would have climbed to 134,290 last month. This is about 250,000 more than reported and can be taken as a rough estimate of Katrina’s independent effect on the overall job market.
Thus, the economy outside the Gulf area was apparently strong enough to offset Katrina’s effect by creating about as much employment as was lost and prevent total nonfarm jobs in September from declining.
Without employment growth, it’s not surprising the national unemployment rate rose from 4.9 percent in August to 5.1 percent in September. The unemployment rise reflected both job weakness in the Gulf area and additions to the labor force from normal population growth.
What is surprising is the nonresponse of the labor force participation rate (proportion of the working-age population in the labor force) to the devastation of Katrina, assuming the measured rate last month is reasonably accurate. With large regional job losses and massive relocation, one might have expected many former Gulf Coast workers to temporarily drop out of the labor force while resettling and struggling to get back on their feet. Yet the September labor force participation rate held steady at 66.2 percent, suggesting a strong desire and need to work. It’s possible, though, that beneath the total an unmeasured drop in the participation rate of Katrina-affected workers was offset by a rise in other workers’ rate.
Looking ahead, the job market is likely to see some further weakness in the next month or two owing to the lingering effects of Katrina and the comparatively milder effects of Rita. Nevertheless, with Gulf area reconstruction picking up and solid growth expected in the general economy, employment should show a distinct improvement in October and later months. To some extent, however, job gains could be constrained by a shortage of skilled workers in the Gulf region.
By early next year, the job market should be back on its pre-Katrina growth path with the economy expanding enough to draw down unused labor and capital resources and further narrow the gap between actual and potential output.
Inflationary pressures emanating from the supply side have intensified recently, mainly in energy, and a resumed further tightening from the demand side seems just around the corner.
The Federal Reserve, rightly so, is looking beyond the economy’s temporary weather-related slowing. Given the central bank’s positive outlook for 2006 and from what Fed officials have hinted, it’s likely the federal funds rate will continue to be raised a quarter-point at each of the next three meetings of the Fed’s Open Market Committee. Allowing for an upcreep in prices, the real fed funds rate should remain relatively low and avoid jolting the economy into stagnation or recession. The Fed’s glide path to a full-employment economy seems about right, with every chance of a soft landing in 2006.
Alfred Tella is former Georgetown University research professor of economics