He looks at data.

Housing starts peaked in January 2006, and then fell steadily for years: January 2006 — housing starts = 2.303 million, unemployment = 4.7% –April 2008 — housing starts = 1.008 million, unemployment = 4.9% –October 2009 — housing starts = 527,000, unemployment = 10.1%

Ouch. I would prefer to use residential construction activity, rather than housing starts (which tend to lead construction spending by several months), but the point remains. Most of the decline in housing construction was behind us by October of 2009.

The other point I would make is that it is unfair to compare almost any variable to employment over the past 2+ years. That is the point of Brad DeLong’s post. In terms of explaining weak employment, the aggregate production function approach falls short also. Or, as Ed Leamer puts it,

the national job markets certain structural problems that have created a mismatch between what employers are looking for and what unemployed workers have to offer. These structural issues include the loss of manufacturing jobs to a variety of “competitors” both technological (robots, microprocessors) and human (foreign workers and recent immigrants willing to work for less) and the housing crisis that continues to jeopardize the construction sector. Unlike in previous recessions, Leamer opines, workers today are not easily returning to the jobs they lost and as a result the economy must find a way to create jobs for millions of workers whose skills lend themselves more suitably to manufacturing and construction.

I think that the behavior of macroeconomic variables in this recession poses some awkward issues for everyone.

1. The big drop in GDP occurs after most of the drop in home construction.
2. The financial indicators, including risk spreads, stock prices, and bank profits, recover fairly nicely, but real GDP does not.
3. Whatever recovery shows up in GDP is not matched by a recovery in employment.
4. A simple linear Phillips Curve of the form inflation = 7.0 – unemployment would say that we should have prices falling at an annual rate of 2 percent now, rather than rising. (Admittedly, this is not a powerful point, because inflation can be measured in various ways, and nobody said that the Philips Curve was linear.)
5. In my recollection (which may be wrong), steep recessions tend to be followed by brisk recoveries. Not so this time.

In short, there is a lot about this recession that I think is puzzling. Sumner has a story for (1) and (5), but it is hard to reconcile with (2) and sidesteps (3) and (4). The recalculation story helps with (2)-(5), but it is hard to reconcile with (1).