For the Record: Stupid Economist Tricks

Nice example of what Bos called a "Marshall McLuhan moment" ("You know nothing of my work," said the Canadian sage, appearing out of nowhere in the ticket line):

How do you do policy right? You do the work like @BharatRamamurti and his staff. That he's more prepared to correctly interpret complicated research than the witness is stunning. https://t.co/aGAACGODfm

— Betsey Stevenson (@BetseyStevenson) September 17, 2020

Fairly clear what happened here: Chris Edwards of the Cato Institute (where he edits Downsizing Government) is testifying before the Congressional Oversight Commission, a new body set up in April to monitor the workings of the CARES act, of which Bharat Ramamurti, just off his service as one of the top economic aides to Elizabeth Warren's presidential campaign is the only so far appointed commissioner, and he is explaining that according to this extremely authoritative paper by Valerie Ramey, government spending does not grow the economy very much, and Ramamurti says well, how about this spot in the paper where it says it does, and Edwards says, approximately, "Homina homina homina—uncertainty!—homina homina," and Ramumurti, says, approximately, "Because I called her up the other night and she said, yeah, that's exactly what she meant."

But that's not how the economics frat boys saw it: they saw Ramamurti as having sprung some kind of dishonorable trap on the witness:

It's clear that he was asking him to repeat what he had already testified to (either previously or in writing), so he could rebut it. I don't think that's baiting, that's what you're meant to do when questioning an expert witness who has offered their expert testimony.

— Betsey Stevenson (@BetseyStevenson) September 18, 2020

It's not a question of whose "fault", it's a question of whether information going into the Congressional Record is accurate or not. Intentionally or not, the witness was offering false testimony.

— Your Strenuous Flu (@Yastreblyansky) September 18, 2020

Edwards: "Ramey says a dollar of federal spending would result in less than a dollar of economic growth." 

Ramamurti: "Later in the paper Ramey says not in a sustained period of extremely low interest rates such as we are now in."

What she told him in the call is identical to the passage from the paper he read aloud. It doesn't differ from the published work the witness cited. The witness cited it incompletely.

— Your Strenuous Flu (@Yastreblyansky) September 18, 2020

I'm sorry, that's simply a lie. Did you listen to the clip? "If interest rates are x for a period of y years then the multiple is in range z" is not an ambiguous formula subject to lots of different interpretations.

— Your Strenuous Flu (@Yastreblyansky) September 18, 2020

Because Ramamurti was quoting from exactly the same paper, of course, as he said, but the attacking economists are somehow unable to hear this. It's a "letter"!

If you find published work he might have referred to, Iet us know: I've already publicly encouraged people to help me find such. I can only report based on what I've found myself, after spending some time looking.

— George Selgin (@GeorgeSelgin) September 18, 2020

Ramamurti explicitly claims that he is reading from the same study Edwards was quoting, not from a "letter". I don't know the literature at all, so I can't help.

— Your Strenuous Flu (@Yastreblyansky) September 18, 2020

Selgin hunted through Ramey's publications and couldn't find anything suggesting what Ramamurti suggested. Well, maybe this one thing...

It's this 2019 JEP paper: https://t.co/7s2MBsKC51. "Some initial explorations," it says, "suggest that government spending multipliers could be higher ... during periods at the zero lower bound of interest rates or wartime."

— George Selgin (@GeorgeSelgin) September 18, 2020

Perhaps I'm still overlooking some more recent and definitive published statement.

— George Selgin (@GeorgeSelgin) September 18, 2020

Actually, that was the paper Edwards had been quoting from, and it just happened to be the paper Ramamurti had been quoting from too, as he said:

It is only after Edwards makes a non-response ("uncertainty...") to this that Ramamurti brings up his phone call with Ramey in which she confirmed what she meant in the passage.

— Your Strenuous Flu (@Yastreblyansky) September 18, 2020

Which you might think would be the end of the story—note that Ramey not only said what Ramamurti said she said, in the paper Edwards quoted

Got it. I missed that. It was the same study I found, eventually!

— George Selgin (@GeorgeSelgin) September 18, 2020

but he decided instead to switch around to litigating the point:

That's from the same paper. Both passages do in fact suggest considerable "uncertainty" about the reliability of the estimates in question. "Some," "initial," "could be" (higher than the usual <1 estimates). Chris had no time to explain what he meant. He might have meant this.

— George Selgin (@GeorgeSelgin) September 18, 2020

Bullshit, he clearly meant to hide it. I could have kept arguing too, because those "initial explorations" have a much broader base than you might think—

The situation is different with respect to periods when interest rates are near the zero lower bound or when monetary policy accommodates government spending increases (such as during World War II in the United States). Numerous New Keynesian DSGE models show that multipliers can be higher than one when monetary policy is constrained by the zero lower bound on interest rates. At the zero lower bound, an increase in government spending provides extra stimulus by increasing expected inflation, which lowers the real interest rate (Farhi and Werning 2016). Calibrated models such as the ones analyzed by Christiano, Eichenbaum, and Rebelo (2011) and Coenen et al. (2012) can produce multipliers that range between 2 and 3 when the period of monetary accommodation is sufficiently long [as seems to be the intention of the Fed, judging from recent statements]. Some recent empirical work has found some evidence of higher multipliers, ranging from 1.5 to 2.5 at the zero lower bound for Japan (Miyamoto, Nguyen, and Sergeyev 2018) and around 1.5 for historical samples in the United States (Ramey and Zubairy 2018). 

but the hell with it, the whole thing has made me too angry, in a somewhat self-righteous way. The thing is, I was right on the basic question—Ramamurti was quoting from the text Edwards claimed was the big evidence on his side, and positively demonstrating Edwards was misreporting it, before he went to the length of talking about his phone call—not because I have any kind of superior knowledge of these issues, of which I indeed have no knowledge at all, but just because I listened to the guys and the content of what they said; and Selgin and co. were wrong, because they didn't—they listened instead for debate points they could use to back up their guy, and as a result actually did not know what had happened in a two-minute video and chased instead through a whole bibliography of irrelevant evidence before coming, with my help, to the facts.

And then Selgin added insult to injury by acting as if he'd found it himself! Oh, hahaha, "it was the same study I found, eventually!" That's what I've been telling you all day, asshole.

I think he still doesn't understand it, because he's so full of himself, and I don't want to talk to him any more.