The National Review decided to run Kudlow’s thoroughly discredited nonsense about output growth, which included these claims:
Well, the marquis employment report for June may have showed “only” 121,000 new nonfarm payroll jobs, below Wall Street expectations. But this leads to another factoid that the mainstream media largely ignores: The household survey of job creation has been booming at a much faster clip than headline corporate payrolls. This survey shows 387,000 new jobs in June, following 288,000 in May. When this last happened in 2003-04 (remember the “jobless recovery” election-year rant of Democrats?), it was corporate payrolls that caught up to the more entrepreneurial household survey – which more accurately records job creation by small-business owner/operators … Wages are rising today, so we know domestic labor markets must be tightening, not softening … The U.S. Bureau of Labor Statistics is recognizing the importance of the small-business-driven household survey, and has suggested averaging household jobs with the corporate payroll survey to get a clearer jobs picture.
Since nominal wages grew by less than inflation – real wages fell. Kudlow is still touting his bogus argument that the household survey is a more reliable measure of employment growth – but as to this last claim, check out this rebuttal:
BLS has always maintained that both surveys are needed to provide a complete picture of the labor market [ed. note: This was not at issue], but BLS has not suggested that one should average the monthly household and payroll survey employment estimates, or the over-the-month changes of these figures.
I wonder if Glenn Reynolds still considers Kudlow to be reliable?
Update: K Harris asked in a comment why the Household Survey increase has exceeded the Payroll Survey increase by almost 1 million. Via Mark Thoma comes Grep Ip explaining the difference in the growth of the two measures of employment growth:
The two surveys often send different signals, but the differences have been especially large lately. Since December, payroll jobs are up 854,000 while household employment is up 1,584,000. Since both surveys estimate total employment from a sample, neither is a perfect tally but economists generally prefer the payroll survey because its sample is far larger – about 400,000 establishments, compared to the household survey’s 60,000 households. In theory, the two could be completely accurate yet come up with widely different job counts, because they define employment differently. In June, these definitional differences accounted for more than the entire difference in the two results.
Of the many definitional differences, the largest, in terms of magnitude, are that the household survey counts the self-employed while the payroll survey does not; and that the payroll survey counts someone with two jobs twice, whereas the household survey counts that person once. Each month, the BLS comes up with a household estimate of employment as using the payroll survey’s definition of employment. In June, that total showed a decline in jobs of 92,000 from May. (You can see the estimates and a technical note on their web site here: http://www.bls.gov/web/ces_cps_trends.pdf1)
Is this positive or negative news for the job market? The answer depends on the reason for the divergence. If it’s because the ranks of the self-employed are growing, the job market is stronger than the payroll survey suggests. If it’s because the number of people with multiple jobs is shrinking, the market is weaker than the household survey shows – all else equal. In June, the message appears to be negative. The number of the self-employed rose 20,000. But the number of multiple job-holders fell 320,000, which means at least that many fewer paychecks from this group. (It may not mean less income: The worker may have gone from two part time jobs to one full time job.) But these two categories swing widely from month to month. In the last six months, the number of the self-employed has risen 346,000 while the number of multiple-job holders has fallen just 75,000. Even using the payroll definition, the household data are still highly volatile; it showed a huge gain of 606,000 jobs in May from April, far more than the payroll survey’s increase of 92,000.
Yet since last August, the two have told almost the same story: Jobs are up by 1.4 million according to the payroll survey, and 1.3 million using the household survey but the payroll definition. The household survey using the broader household definition is up by 1.9 million, which means about 600,000 extra jobs that can’t be explained by definitional differences. It may be a sign, albeit an inconclusive one, that job creation is stronger than the payroll survey suggests.