WSJ: “The CEA memo uses publicly available unemployment insurance records to calculate that employment from March 2003 through December 2003 grew by 288,000, or 32,000 per month, more than previously published BLS estimates. The memo says “it is tempting” to extrapolate the monthly figure out to March 2004, producing a total increase of 384,000. But it downplays the higher figure, warning that employment in the first three months of the year could well be revised down, not up, citing other data revisions that tilt in that direction.
Even with positive revisions, Democrats probably will be able to attack Mr. Bush as the first president to oversee no net job creation since Herbert Hoover. Republicans have countered that the Bureau of Labor Statistics’ household survey shows employment actually up 1.9 million, or 1.4%, under Mr. Bush. But a study published earlier this year by the Federal Reserve Bank of Cleveland finds the household survey, examined more closely, tells a much less positive story.
The staff study found that when the most reliable part of the household survey is compared with the payroll survey, “both measures … show a surprisingly similar picture of the weak labor-market performance that has prevailed during this recovery relative to previous business cycle periods.”
What other explanations might exist for this divergence? I continue to harp on the post-bubble environment as the single biggest factor. Regardless of which survey you look at, the lag is there:
Cleveland Fed economists Mark Schweitzer and Guhan Venkatu note in their study that the household survey’s employment total can be distorted by problems in extrapolating from a sample of 60,000 households to the total population, because of uncertainty surrounding population estimates.
The economists look instead at the percentage of the working-age population that is employed, which they write is “more informative and less problematic” because it factors out population. The authors say that in the nine post-World War II recoveries prior to the most recent, this “employment to population ratio” fell on average 1.5 percentage points in the 18 months after recession began. By the three-year mark, it was down less than half a point.
After the most recent recession began in 2001, the ratio tracked the postwar average for the first 18 months, but then continued to decline. By the three-year mark it was down 2.2 percentage points from the peak.
“This picture is strikingly similar” to the poor performance of payroll employment relative to previous recoveries, they write. The authors estimate payroll employment rose 3.7% in the first three years of the nine previous business cycles, but is still down 1.5% in the latest.
The chart reveals all!
Political Debate Over Jobs Intensifies
Bush Team Hopes for Lift From Friday Payroll Data;
Revisions to Get Attention
WALL STREET JOURNAL, October 5, 2004