Sumner is one of those Right wing economists who tries to claim that some very simple things that exist, don't exist at all. It's fascinating to me that when we discuss tax policy, the Right wing pundits and think tanks all come out with apologetic op-eds claiming that even to call someone rich is out of bounds, so when ever they discuss the question of wealth distribution they call the rich, "the rich" as if it's debatable whether anyone in the world is or has ever been in a position to be accurately described as rich.
Another word that guys like Sumner try to pretend is somehow problematic is "bubble", so there was not bubble in 2004-2006, and the only problem was that the Fed let NGDP fall. In his latest, Sumner concludes there was no bubble based on a highly selective analysis of housing starts and the unemployment rate between 2006-2009. He looks at housing starts vs. housing completions and then the unemployment rate and that tears it. Let's listen to his oh so facile analysis:
First the numbers that cinch it:
"an. 2006: starts = 2,303,000, completions = 2,058,000, average = 2,180,000, U-rate = 4.7%"
"April 2008: starts = 1,008,000, completions =1,014,000, average = 1,011,000, U-rate = 4.9%"
"October 2009: starts = 527,000, completions = 745,000, average = 636,000. U-rate = 10.1%"
Really not clear what this is supposed to prove. In 2009, starts and completions where way down as was the average of starts and completions and the unemployment rate had swelled to 10.1% and this shows there was no bubble?
Obviously he confuses the correlation of the bubble and high unemployment. No one thinks that during a bubble, certainly not a housing bubble' you're going to see unemployment spike, to the contrary you'd expect to see it low during the bubble. The idea is that bubbles eventually pop and when they do, the fallout will be a lot worse than otherwise would have been. If there is a bubble today, it will at some point lead to high unemployment tomorrow. To compare current housing starts and completions in 2006 to the unemployment rate of the time is comparing apples to oranges.
The correct comparison would be housing starts and completions in 2006 and unemployment in 2009. After wholly misconstruing the analysis, Sumner can declare with typical facileness that:
"If the great housing construction crash of January 2006 to April 2008 didn’t “drive the economy too far from its full employment path,” I think we can safely assume that no housing crash will ever cause a recession in the US"
"Too bad the Fed let NGDP fall in late 2008."
Listen, no one believes in the virtue of simplicity more than me, but often the way Sumner and the rest of the Right wing apologists-Cochran, Lucas, et al.-use simplicity mangles it into a vice. Is it not obvious that Sumner's whole NGDP is such a case of overstating his case by oversimplification?