Tuesday, November 8, 2011

Why You Should Pay Attention to Payroll Employment Revisions

Ed Yardeni writes:
Pay no attention to the first estimate of payroll employment provided monthly by the Bureau of Labor Statistics (BLS). I have been giving this advice for many years, as I’ve observed that during economic expansions the data tend to be revised higher, and lower during recessions. I’ve argued that in assessing the strength or weakness of the economy, more weight should be given to the previous two months of data and their revisions than to the first preliminary estimate of the latest month. The most recent revisions are certainly encouraging, as are several of the other labor market indicators:
This falls in line with my argument that Keynesian economists and econometricians don't have a theory about the business cycle that matches reality and that therefore they are slow in detecting turns in the economy. They are too bullish as the economy is weakening and too bearish as the economy is improving. Since the BLS is mostly comprised of Keynesians and econometricians (though not all, there are a few EPJ readers), it is no surprise that their early estimates are off and directionally against a changing trend.

Here is a chart Yardeni has put together showing the revisions and the direction of the revisions:

And here's Yardeni's summary on recent revisions:
Revisions find more jobs. Nonfarm payrolls were revised upward by 102,000 during August and September. During the third quarter, payroll employment rose 129,700 per month on average, up from 96,700 during Q2. October’s preliminary (and unreliable) gain was 80,000. Private sector payrolls were revised up by 84,000 during August and September, and rose 145,000 per month on average during Q3, while October’s gain was 104,000.

Over the past 12 months through September, revisions added 274,000 (based on first estimates) to total payrolls, which rose 1.59 million over this period (see chart above). Private sector employment rose 1.82 million over the past 12 months through October. This jibes with the ADP survey of private sector employment, which tends to be revised much less than the BLS data, and is up 1.74 million y/y through Octobe
Yardeni also has a nice chart of one of my favorite employment indexes, the Monster Employment Index of Online Job Ads:

This index rose to 151 during October, the highest reading since September 2008. Leading the way is the transportation and warehouse industry. Not far behind is the retail industry. The indexes for both industries have rebounded back to their record highs of 2007. Lagging, but still making decent comebacks, are the manufacturing and information technology indexes. The big laggard is the index for the finance and insurance industry.

Bottom line: The economy is recovering to a much greater extent than most realize.


  1. You want proof of how clueless economists are, just head on over to Murphy's blog where the Austrians are trying to re-teach everyone the law of diminishing marginal utility. You think I am joking? I only wish....

    Sorry to post a link to another blog, but this is seriously getting ridiculous.

  2. Look at the charts, especially the 12-month
    rolling revisions- they're well below prior experience. 1.75 million new private sector jobs in past 12 months-that's under 150K a month and that's a very weak performance in the current context. Real bottom line, the economy is experiencing the worst business expansion
    cycle in decades. Whether this is 'much greater'
    than what 'most expect' I don't know, but the actual results are bad.