Should-Read: IMHO, the smart PGL does good here...
And, J.W. Mason, please don't say that Coibion et al.'s state-of-the-art potential output estimate "gives a very similar estimate for the output gap as simply looking at the pre-2008 forecasts or extrapolating from the pre-2008 trend". That's not the way to gain a reputation—at least, not they way to get any reputation that you would like to have.
The Blanchard-Quah concept potential output estimates being made by Coibion, Gorodnichenko, and Ulate http://delong.typepad.com/w23580.pdf do not look to me "very similar estimate for the output gap as simply looking at the pre-2008 forecasts or extrapolating from the pre-2008 trend". Look at the gap between the current estimate—the purple line—and the red or dot-dash green line that are the 2007 and 2009 vintage BQ forecasts. Look at the thick blue line that is (my drawing) of the upper hull of real output in this millennium a la Menzie Chinn:
Yes, BQ gives a large current output gap. No, BQ as applied by Coibion et al. does not give a "very similar estimate for the output gap as simply looking at the pre-2008 forecasts or extrapolating from the pre-2008 trend". You should not say it does.
PGL: The Output Gap per the Gerald Friedman Defenders: "Menzie Chinn back on February 20, 2016 had some fun with the defenders of that awful paper by Gerald Friedman... http://econospeak.blogspot.com.br/2017/08/the-output-gap-per-gerald-friedman.html
...the trend line extrapolated from 1984-2007 implies that the output gap as of 2015Q4 is... -18%... I want to stress that estimating potential GDP and the output gap is a difficult task...
It is a difficult task and perhaps the CBO estimate of the gap is understating it. But to pretend the potential GDP would continue to grow by something akin to 3.4% from 2000 to 2015 is just absurd. Well it seems this crowd has “updated” their trend line analysis. J. W. Mason writes:
This alternative measure [from Coibion et al.] gives a very similar estimate for the output gap as simply looking at the pre-2008 forecasts or extrapolating from the pre-2008 trend...
The ‘alternative measure’ comes from an interesting new paper... [that] suggests that the output gap may be as large as 10%. [But] Mason’s back flip is not justified...
My view as to how much slack there is and the desirability of expanding aggregate demand is and always has been: we don't know how much slack there is, expanding aggregate demand to see is good policy.
My guess? That BQ produces estimated output gaps that are larger than the real thing. Applying Okun's Law to the prime-age employment-to-population ratio gives us 4%-points of slack. That's the hill I would die on. But my real answer is: "we don't know, and we should pursue expansionary policies to find out".
We shouldn't promise unicorns and rainbows: it's highly unprofessional to say that we have even some confidence that sufficient demand expansion would get us back to the pre-2008 trend; and it's wrong to cite Coibion et al. in support of that claim. And CBO's current forecast of potential output growth is a reasonable one in the absence of a major policy shift to major aggregate demand expansion.
And: Mason: don't write: "I don't really want to engage with this guy... but if there's a lone crank in this discussion it ain't me..."
And don't write that your claim that potential output today is well estimated by "simply looking at the pre-2008 forecasts or extrapolating from the pre-2008 trend" has been "reviewed by Obama CEA chair Jason Furman and Laurence Ball... presented... to a group including Furman, Dean Baker, Josh Bivens, Lawrence Mishel and a number of other people with solid macro credentials... [with a] positive response..."
Your Roosevelt paper http://rooseveltinstitute.org/wp-content/uploads/2017/07/Monetary-Policy-Report-070617-2.pdf is much better than your blog post http://jwmason.org/slackwire/what-recovery-the-conversation-continues/ or your comment http://econospeak.blogspot.com/2017/08/the-output-gap-per-gerald-friedman.html?showComment=1501729914146#c1316241742813017907. Stick to the Roosevelt paper.