Friday’s jobs report brought welcome news that the unemployment rate had fallen to 7.8 percent in September, a surprisingly large decline given the mediocre 114,000 nonfarm jobs that were added. The report showed that about twice as many people got hired as entered the labor force, causing the September unemployment rate to fall.
Economists recognize the inherent volatility of monthly statistics and are particularly unmoved by the unemployment rate in any particular release, given the degree of variability in the data. They tend to focus more on the trend in job creation over several months.
However, some particularly uninformed and addled comments surfaced after the latest report. In particular, former GE chairman Jack Welch leveled accusations of data manipulation against the Obama administration. Given the surprising lack of ridicule with which Welch’s comments were greeted, a primer in how the data are reported seems in order.
On the first Friday of each month, the Bureau of Labor Statistics releases results of two separate surveys. The Current Employer Survey (CES) polls 141,000 businesses and government agencies to collect information on jobs created or eliminated. This so-called “establishment survey” is the source of the nonfarm payroll statistics to which economists tend to pay the most attention.
The unemployment rate is a product of the Current Population Survey (CPS), also known as the “household survey.” Each month, 2,200 professional Census Bureau data collectors contact 60,000 separate households with a series of questions aimed at capturing the employment status of residents aged 16 and over. The monthly household survey has been conducted continuously since 1940 by experienced career civil servants who are not political appointees. In fact, to paraphrase one knowledgeable economist, these are quintessential “nerds with pocket protectors” who bristle at any suggestion of manipulation.
Results of the surveys are tabulated, seasonally adjusted, and summarized by small groups of professional BLS economists sequestered behind locked doors and cut off from outside communication. Advance copies of the results are delivered to the Federal Reserve Board and to the White House the day before release, and are communicated to the Labor Secretary on the morning of publication. Financial news media are given embargoed copies at 7:45 a.m. on Friday and are allowed to report the results at exactly 8:30 a.m. A copy of the release is simultaneously posted on the BLS website.
Americans are particularly fortunate to have a plethora of public data at our fingertips to inform a vigorous debate over budget and policy priorities. While a legitimate disagreement over specific methodologies is entirely appropriate, the integrity of the actual data is not in question. Any suggestion to the contrary should serve to disqualify the discussant from further consideration.
Responsible commentators suggest that next month’s data could well bring revisions to the surprising September result, something which would be entirely consistent with historical precedent given the noise in the monthly data.
Meanwhile, Jack Welch later amended his comments to suggest that a question mark be added. At the very least.
Christopher A. Hopkins CFA, is a vice president at Barnett & Co.