The stock market took a huge plunge Friday
partly because of disturbing unemployment data. But one analyst, a guy who likes to follow arcane things like
the price of prepainted steel coils and insulation, argues that the government's information is misleading. He says the Feds are counting wrong.

I got a "D" in college economics, so the observations of the economist, Kenneth Simonson, are a little beyond me. But the analysis he released today after the stock market closed is interesting, at least the part I could understand.

Here's what he had to say:

1) The June 6 employment report from the Bureau of Labor Statistics (BLS) indicated that construction employment fell 34,000 in May, seasonally adjusted, and 386,000 (-5.1%) over the past 12 months. Reportedly, the two residential construction categories (residential building and specialty trade contractors) shed 316,000 jobs (-9.6%) since May 2007, while nonresidential construction (building, specialty trade and heavy & civil engineering) lost 70,000 jobs (-1.6%).
However, the Census Bureau reported on June 2 that residential construction spending declined 21% in the 12 months through April 2008. It is very unlikely that homebuilders and residential specialty trade contractors have reduced their headcounts by less than 10% when they are performing 21% less work, especially since building permits and builders' sentiment surveys consistently show a further steep decline in activity. The Census figures also show a 12% increase in nonresidential construction spending from April 2007 to April 2008, a number that would be hard to achieve with a reduction in employees, as counted by the BLS data.
It appears likely that many of the workers being counted as "residential" specialty trade contractors are actually working on nonresidential projects, such as wiring, wallboard installation and concrete finishing for schools, offices, hotels, etc. Indeed, many general contractors have reported that subcontractors who formerly worked on housing are now bidding for nonresidential work. Such a misclassification could occur if a company listed its industry code as residential when it entered the BLS data base but it is now doing nonresidential work and does not change its code.
If residential construction employment is assumed to have fallen 21% rather than 10%, that implies a further drop in residential employment of 360,000 workers. If those employees are added to the nonresidential total, nonresidential employment shows an increase from May 2007 to May 2008 of 290,000 (7%), not a decrease of 1.6%. A 7% increase is much more consistent with the Census estimate of a 12% increase in nonresidential spending, especially since that spending figure includes substantial material cost increases.
The employment report also showed that average hourly earnings for construction industry employees rose 3.7% in the past 12 months, compared to 3.5% for all private sector production or nonsupervisory workers. For construction wages to be rising faster than overall wages when construction employment is dropping (-5.1%) and total private employment is flat is improbable. But such a wage increase is consistent with rising employment in the more highly skilled nonresidential trades and falling employment in the less skilled residential trades.
The Pew Hispanic Center reported on June 4 that 250,000 Hispanic workers lost jobs in construction in the past year; anecdotal reports suggest the bulk of those workers were in residential new construction and improvements. They also tended to be lower-paid than nonresidential employees or even than the median pay for residential construction workers. If they lost jobs disproportionately, the median wage would rise.