Public opinion polls indicate that the economy may be the overriding issue in the presidential election in November. Job creation, or the lack of it, is already being given great prominence in the strident and distressingly early electioneering already under way.

According to those who favor a second term for the current Administration, the economy hasn't done too badly, considering that it has had to contend with an inherited and unavoidable recession and the impact of the Sept. 11, 2001, terrorist attacks and the subsequent wars.

They also stress that the economy has been growing since November 2001 and that the job turnaround now getting under way has been delayed by spectacular economy-boosting gains in productivity. A survey by Monster Worldwide does show a steady increase in online job recruitment activity since last October.

Those who favor change in Washington, D.C., charge that job performance has been dismal. They relish pointing out that the current Administration will likely be the first since Herbert Hoover's to witness an overall four-year drop in payrolls.

The statistical base for any analysis of the employment situation is two monthly surveys by the Bureau of Labor Statistics (BLS). One, the household survey, is of the residents of a sample of 60,000 households. It provides data on the numbers employed and unemployed. It also measures how long those who are without work and are looking for it have been unemployed.

The other, the payroll survey, is of a sample of 160,000 businesses and government agencies covering about 400,000 establishments. It does not measure unemployment.

The payroll survey is regarded as the more reliable measure of employment. It is benchmarked annually against employment counts derived primarily from unemployment insurance tax records--a definitive measure. The household survey cannot be directly benchmarked against independently derived data. But it is adjusted annually to underlying census estimates. These adjustments can be quite large. For instance, an adjustment resulted in a boost in employment of almost 1 million between December 2002 and January 2003.

A significant difference between the surveys is that the payroll count excludes farm workers, the self-employed, and much casual day labor. However, over the long term, employment data from the two surveys parallel each other, with the household count several million higher than the payroll count.

The bare-bones data show the following:

• An increase in employment, as measured by household survey data modified by BLS to smooth out the annual adjustments, of from 137.7 million in December 2000 to 138.1 million in December 2003. This is an increase of just over 100,000 per year, far below the annual increase of 1.2 million in the workforce over the period.
  
• A hiatus in payroll growth unprecedented since the Great Depression. This March, at 130.5 million, the payroll numbers were still 2 million lower than they were three years ago. But they are up from the low of 129.8 million last August and showed a sharp upsurge of 300,000 last month.
  
• A median amount of time of just over 20 weeks for those unemployed today to find a new job. This is the longest it has been for the past 50 years except for during the sharp recession of 1983.

With numbers like these, spinning is harder for Bush Administration advocates. They have been stressing the household survey data, which at least show the modest gain since 2000. For a variety of reasons that are at times convoluted, they discount the payroll data, claiming that the persistent deficit is due largely to the way it is counted. Ironically, the recent uptick in job creation has come in payrolls. They are up by more than 500,000 for the first three months of this year, while total employment, as measured by the household survey, has dropped by 200,000.

Critics of the Administration can't resist comparing the 23 million increase in payrolls over the two Clinton Administrations with the 2 million drop since. However, they charge, somewhat disingenuously, that if the boom of the 1990s had been kept going, payrolls would be 7 million higher than they are today. However, every boom causes economic strains that eventually have to be relieved with a slowdown. The 1990s' upsurge was already the longest and strongest in U.S. history when it ended in early 2001, and it had generated some economic bubbles that were already bursting.

Chemists have been swept along by all of this. Last year's annual ACS survey of the salaries and employment status of its members in the domestic workforce recorded the highest level of unemployment in its more than 30-year history (C&EN, Aug. 4, 2003, page 37).

The society's latest annual survey of new chemistry graduates (see page 51) reveals that the job market was still softening last fall and indicates some salary weakness. The volume of classified employment ads in C&EN, a strong current indicator, has been tumbling for three years and has not yet clearly turned around.

However, the faltering job creation of the past three years has not had the devastating impact on chemists that the economic downturn of the early '70s had. At that time, Ph.D. graduates were apocryphally, if not literally, driving cabs, and ACS was in an uproar over what it should do about it.

The hope has to be that whoever wins the election this year will appreciate that doing what he can to foster job creation is as important as fostering wealth creation and that the two are not always hand in glove.