The Washington Post has a piece puzzling over the fact that:
"Builders started work on 27.7 percent more homes in February than they did a year earlier. Yet the number of construction jobs in the United States was only 2.9 percent higher, year-over-year."
The Post turned to analysts at Goldman Sachs who concluded that the answer was labor hoarding. They make a case that firms are changing the length of the workweek to meet the increased demand for labor rather than adding more workers. This one doesn't fly.
First, residential construction is a comparatively low-paying sector with casual labor relations. This is not General Motors with union contracts that make layoffs difficult. In other words, it is not a sector where we would expect to see a lot of labor hoarding.
The data on hours shown in the piece also do not support the labor hoarding story. While the average workweek has increased by roughly 3 hours since the trough of the downturn in 2009, it is up by only about 0.5 hours since 2011, which means that it would be equivalent to an increase in employment of less than 2 percent. That will not fill much of the gap identified in the piece.
So, what's the real story? First, total construction is up by much less than residential construction. The Commerce Department reported that total nominal construction was 7.1 percent higher in January of 2013 than January of 2012. In real terms this would be a rise of around 5.0 percent, not too different from the increase in employment.
The other big issue is that many of the workers employed in residential construction are undocumented and may not show up in the payroll data. In fact, there was a sharp decline in residential construction in 2006 and 2007 even as employment in construction was still growing. From its peak in 2005 to the end of 2007 housing starts fell by almost 40 percent, while construction employment was little changed. Given this history, there is no reason to expect a big upturn in employment in response to the relatively small rise in starts that we have seen in the last year.
Click for larger version
(Only one link allowed per comment)