by Calculated Risk on 4/24/2007 01:22:00 PM

In a research note titled "US Views: On Track?", Jan Hatzius, Chief US Economist at Goldman, Sachs & Co. commented today on residential construction employment.

"We expect ... evidence of labor market slowing over the next few months. ... our expectation [is for] a large decline in construction employment ..."
Jan Hatzius, April 24, 2007

Note: Excerpts used with permission.

Click on graph for larger image.

This graph shows starts, completions and residential construction employment. (starts are shifted 6 months into the future). Completions follow starts, and employment usually tracks completions.

Based on historical correlations, it is reasonable to expect BLS reported residential construction employment to follow Starts and Completions "off the cliff" soon. One of the current mysteries is why residential construction employment (the blue line) is holding up so well.

Hatzius doesn't believe the problem has to do with the mis-classification of employees (some have suggested that the strength in non-residential construction might be masking weakness in residential construction employment):

"... the mis-classification hypothesis can't explain why overall construction employment is flat while overall construction activity ... has been falling [significantly]."

And what about undocumented illegal immigrants?

... does the layoff of undocumented immigrants who don't show up in the statistics "insulate" the payroll data, i.e. do the illegals lose their jobs so the regular employees can stay on the payroll? ... I don't think the story ultimately holds up. First, it can't explain why residential payroll employment did track activity very closely on the way up, but (so far) not on the way down. Second, while remittances to Mexico -- one of the few proxies for undocumented immigrant employment -- have fallen slightly over the past year, the pattern broadly tracks residential construction employment (which after all is also down, just not as much as we had expected). Third, the household survey of employment, which may pick up a larger share of immigrant employment because it includes independent contractors, shows an even stronger performance of construction employment than the establishment survey over the past year (though there is no res/nonres breakdown).

Update: On the following remittance data from the WSJ, Hatzius has sent me a graph of the growth rate (using a 3 month moving average), and the slow down in remittance does appear gradual. From the WSJ today: Latin America Feels Pain Of U.S. Housing Slump

Monthly remittances from the U.S. to Mexico have dropped every month since their peak of $2.6 billion in May 2006 -- shortly before new-home construction in the U.S. plunged. In February 2007, the latest month for which data are available, remittances to Mexico had slowed to $1.7 billion.

Mexico, Latin America's remittance leader, may be a leading indicator of a trend unfolding across the continent. In a recent study of 15 Latin American economies tracked by BCP Securities of Greenwich, Conn., all but three showed better than a 90% correlation between the ebb and flow of U.S. housing starts and the swelling and shrinkage of remittances as recorded by the nations' central banks.

So what is the answer to the residential construction riddle? In an earlier note, Hatzius suggested there might be some "hoarding" of employees, as many employers anticipate a turnaround in housing. From Roubini: Explaining the Mystery of Why Housing Jobs Have Not Fallen Much...and the Worsening Housing Recession...

The [hoarding] argument, presented by Jan Hatzius, the excellent U.S. economist for Goldman Sachs, is that home builders decided not to fire workers when housing starts started to fall last year because they were expecting a housing recovery this year. So they hoarded labor. Then the observed fall in housing construction productivity that we see in the data since 2006 is directly related to this labor hoarding. However, he argues that, if as likely, the recovery of housing does not occur this spring the home builders will have to start laying off such workers. So we will soon see the effects of this on the labor market.

Another possibility is that the BLS is missing a turning point in the residential construction employment area. From the BLS:

The most significant potential drawback ... is that time series modeling assumes a predictable continuation of historical patterns and relationships and therefore is likely to have some difficulty producing reliable estimates at economic turning points or during periods when there are sudden changes in trend.

Whatever the reason, Jan Hatzius expects BLS reported residential construction employment to decline significantly over the next few months. So do I.