Tarnished Silver
krugman.blogs.nytimes.com
krugman.blogs.nytimes.com
> But data never tell a story on their own. They need to be viewed through the lens of some kind of model, and it’s very important to do your best to get a good model. And that usually means turning to experts in whatever field you’re addressing.
His first two sentences are spot on - data must be matched with a story (I might not use the word "story" here, but the general principle is right).
The amusing part though is that he assumes that Nate Silver isn't an expert at this point. I've been following 538 since before 2008, and the quality of analysis from a non-statistical level was consistently far higher than almost any other publication (except perhaps the Economist).
If anything the quality level once 538 signed with the Times decreased slightly, but only because a number of the other contributors (Andrew Gelman, etc.) stopped contributing at that time.
[0] Krugman hit an impressive low in January with this column - a rather vapid dismissal of the Silk Road as (and I quote) "libertarian derp", devoid of any actual analysis or supporting arguments whatsoever: http://krugman.blogs.nytimes.com/2014/01/19/derp-pirate-robe...
Are you kidding? I love Paul Krugman. He's my favorite comedian!
Krugman isn't disagreeing that Silver is good at political analysis. In fact, Krugman has been a huge supporter of Silver and his work, because (at least to my eyes) Krugman identifies with him. Keep reading the post.
"In that case, he pitted his statistical approach against campaign-narrative pundits, who turned out to know approximately nothing. What he seems to have concluded is that there are no experts anywhere, that a smart data analyst can and should ignore all that. But not all fields are like that — in fact, even political analysis isn’t like that, if you talk to political scientists instead of political reporters."
If you've read the new 538, there are a great deal of non-political analysis stories - covering, for example, climatology and economics - that just have it all wrong, for the very reasons Krugman is talking about. Krugman is not the only one saying this.
>Krugman hit an impressive low in January with this column - a rather vapid dismissal of the Silk Road as (and I quote) "libertarian derp", devoid of any actual analysis or supporting argument
That's not a column, it's a blog post. Krugman posts lots of unserious stuff on his blog, like his undying love for Arcade Fire. He's spent plenty of other posts giving as much thought he as a Keynesian who readily admits he's not technologically saavy to Bitcoin-related matters should. If you want serious analysis, read his economics-related posts. If you care about what Krugman the man feels about something (and I don't know why you should, unless you know him personally), read the others.
I'm not so sure. He did predict that quantitative easing would not cause inflation; but the reason QE has not caused inflation is that banks are not making loans; they are letting $2.4 trillion of QE money sit in their accounts at the Fed instead of loaning it out.
Krugman most certainly did not predict that outcome; he said QE would cause a recovery in the loan market. It hasn't. Getting one prediction right by accident because you got another prediction egregiously wrong doesn't make you a good predictor.
Can you provide a citation for that claim? Because my recollection is that he said QE would have limited effects. I can't find much supporting this memory right now, aside from this entry about QE2[1]. He definitely supported QE, but I don't think he ever argued it would fix everything by itself. He's consistently argued that the Fed can't fix things by itself.
On the other hand, here's this entry where he says that, contrary to what you suggest, the whole point of QE is to cause inflation [2].
[1] http://krugman.blogs.nytimes.com/2010/10/23/how-to-think-abo... [2] http://krugman.blogs.nytimes.com/2009/03/20/fiscal-aspects-o...
You're correct, he did say that back in 2008. See Amezarak's post in response to me upthread.
he says that, contrary to what you suggest, the whole point of QE is to cause inflation
I think he meant that the Fed thought it would cause inflation. But he has said a number of times that if the economy is in a "liquidity trap", printing more money will not cause inflation. (QE is one way of effectively printing more money.)
http://krugman.blogs.nytimes.com/2013/04/28/monetarism-falls...
The results are in, Krugman was wrong.
http://econlog.econlib.org/archives/2014/01/the_parrot_is_s....
Actually Krugman has been getting the consequences of austerity quite wrong. For example, the sequester failed to cost 700,000 jobs:
http://www.nytimes.com/2013/02/22/opinion/krugman-sequester-...
Silver built his reputation by taking the results of professional pollsters seriously, rather than attempting to spin them as most political reporters do. It was not a high bar to clear.
If he's going to report on climate change, or indeed any area of scientific study, then he needs to start with what actual scientists are reporting, just like he started his political reporting with what actual polls were saying.
http://nipccreport.org/reports/ccr2a/ccr2physicalscience.htm...
Also, Pielke's article referenced the IPCC AR5's own conclusions about the lack of evidence for an increase in extreme weather events.
So I'm behind the times. Why are most of the comments on this article pulling out all well-worn Hacker News tropes for dismissing the article and/or making the field slippery enough that I can ignore what Krugman says?
As far as I know, Krugman and Silver are both regarded as real smart dudes in their respective fields (both with actual political hacks ready to paint them as hacks), and nobody is expected to bat 1.000 so this looks like a "hey kid new to these waters, be careful there be dragons."
So is the drama over a politically sensitive topic? I'm assuming. Is the narrative about the issue coming out of 538 more pleasing to the ears of entrepreneurial self-styled "hackers" than the one from Krugman? I bet so! Now that I've posted this, I'll go catch up on the drama to find out just how wrong I am.
Doing your own analysis doesn't mean you believe there are no experts. It means you believe experts are sometimes wrong and there's value in doing original analysis from the source data.
I love that about 538. Most news sources are writing based on the same small set of sources and quotes. Even if I don't agree with 538's analysis, it's useful to get the perspective of a critical outsider armed with data.
Krugman isn't criticizing Silver for doing his own analysis, he's criticizing the lack of judgement he displays is hiring Roger Pielke, Jr., -- who is neither a neutral data scientist nor a climatologist, but a political scientist who has long been selling a very specific position regarding climate change -- to do his analysis for him on the issue of climate change, and then publishing his report with the specific flaws that critics have identified in Pielke's piece, and suggesting that the process that allowed all that to happen was a result of Siver's disregard for expertise in the domain.
Personally I think that Krugman is attacking Silver for the wrong reason -- I don't think that Silver's problem is a lack of respect of real expertise in any domain (and in his own work, while he was conducted his own analysis, he has also demonstrated respect for -- without unquestioning deference to -- domain experts), I think its more likely that if there is a problem its just a simple failure of judgement regarding someone else's abilities and interests.
He does not rehash them for brevity.
This statement from Krugman, to me, flies in the face of scientific thought - "you should be very wary if your analysis runs counter to what a lot of professionals say." Not at all true - the "lot of professionals" should be wary if a scientific study runs counter to their analysis.
The best value of statistical wizardry is when it works alongside domain experts.
While it's superficially true that this expertise can be used to enshrine unscientific dogmas in the trappings of science, to reject expertise period is to reject the role that knowledge can and must play in forming an understanding of our world. Data do not exist in a vacuum, but are collected in actual experiments, the complexities of which must be understood to gain insight from that data in a way that reflects reality. Not being wary of (or even intentionally misconstruing) how data is collected and represented leads to mistakes like normalizing away trends, then claiming those trends don't exist; this is precisely what Pielke Jr. did, and what Krugman was criticizing.
You are arguing against a straw man. Neither Silver nor Pielke claimed it wasn't.
...to reject expertise period is to reject the role that knowledge can and must play in forming an understanding of our world.
No, rejecting "expertise" is simply rejecting the idea that you must outsource your thinking to others.
Nate Silver believes you should be purely frequentist viewpoint of the data first.
Krugman is pushing the Bayesian viewpoint. Given experts' assumptions (that have been built up from examining lots of prior data with critical thinking), let's examine this data.
This is a battle that is being fought on many other stages than Krugman's blog.
Does he? A recent column on 538 (not written by Silver himself, true) talked about using Bayesian methods to help find the missing airliner:
http://fivethirtyeight.com/features/how-statisticians-could-...
Krugman is probably worried about what pure stats will say about the predictions of most economists.
But the reason it hasn't is that banks are not making loans with the new money; they are letting $2.4 trillion of it sit in their accounts at the Fed. Krugman most certainly did not predict that; he said QE would help restart the loan market. So he got one prediction right only by accident, as a result of getting another prediction egregiously wrong.
Krugman, 2008: http://krugman.blogs.nytimes.com/2008/09/22/the-humbling-of-...
2009: http://krugman.blogs.nytimes.com/2009/12/14/followup-on-samu...
2013: http://krugman.blogs.nytimes.com/2013/08/14/hawks-doves-and-...
I don't know what it is about Krugman, but he seems to have his views misattributed more than any other commentator.
http://krugman.blogs.nytimes.com/2014/03/19/nobody-could-hav...
You're right; I had forgotten that he pointed out that QE didn't help Japan's loan market at all.
Krugman's attitude seems to have mostly been "if we can't have fiscal policy that would actually solve the problem, then I guess we'll have to hope that QE does something but I doubt it."
Yes, he often talks about "forward guidance" as his preferred method. I didn't mean to say he was in favor of QE in preference to other methods.
Personally, I'm as skeptical of "forward guidance" as I am of QE, mainly because the Fed has, for quite some time now, been trying to double-promise that it won't raise interest rates for a while even if the economy starts recovering. The problem is that nobody appears to believe them.
But it's not; it's a content website that sells ads, and according to Silver, aspires to deliver truthful reporting. Hirng Roger Pielke to draw naive correlations is neither research nor truthful reporting.
Krugman's mentions "But data never tell a story on their own.", and does the polar opposite, tries to tell a story with no data, and that's the wrong kind of story for what he was aiming for.
Using Krugman's own words, his post "looks like something between a disappointment and a disaster."
However, Krugman often takes it (as in this case) to the opposite extreme – like many economists, finding a model, then examining the data and interpreting it to fit that model. This sort of confirmation bias and narrative-based news is what Nate Silver is trying to oppose, not the lack of a narrative or model whatsoever.
There's an interesting experiment from the UK Daily Mirror (a broadly left-leaning tabloid newspaper) called ampp3d: a website which tries to use facts and infographics to tell a story. The editorial tone varies quite a bit from story to story, but tends towards the informal and light. I like aspects of their approach and can see the potential but I actually wish they were more rigorous in picking their sources.
Here's an example where they try and unravel claims about welfare benefits:
http://ampp3d.mirror.co.uk/2014/02/04/wed-rather-have-big-be...
> We've learned and struggled for a few years here figuring out how to make a decent phone. PC guys are not going to just figure this out. They're not going to just walk in.
"Senator Sessions quoted from testimony before the environment and Public Works Committee the previous July by Dr. Roger Pielke.... Specifically, the Senator read the following passages from Dr. Pielke’s written testimony:
"'It is misleading, and just plain incorrect, to claim that disasters associated with hurricanes, tornadoes, floods or droughts have increased on climate timescales either in the United States or globally. Drought has “for the most part, become shorter, less, frequent, and cover a smaller portion of the U.S. over the last century”....'
"Footnotes in the testimony attribute the [statement] in quotation marks [...] to the US Climate Change Science Program’s 2008 report on extremes in North America....
"...[A]ny reference to the CCSP 2008 report in this context should include not just the sentence highlighted in Dr. Pielke’s testimony but also the sentence that follows immediately in the relevant passage from that document and which relates specifically to the American West. Here are the two sentences in their entirety (http://downloads.globalchange.gov/sap/sap3- 3/Brochure-CCSP-3-3.pdf):
"'Similarly, long-term trends (1925-2003) of hydrologic droughts based on model derived soil moisture and runoff show that droughts have, for the most part, become shorter, less frequent, and cover a smaller portion of the U.S. over the last century (Andreadis and Lettenmaier, 2006). The main exception is the Southwest and parts of the interior of the West, where increased temperature has led to rising drought trends (Groisman et al., 2004; Andreadis and Lettenmaier, 2006).'"
Yes, Pielke argued ¬P by carefully picking a quote immediately next to a statement of P.
Extra credit for the mention of "Dr. Roy Spencer of the University of Alabama", who is the smarmiest goof-ball I've ever heard present. Really, no one is a bigger fan of contrarian scientists than I am; one of the best things Einstein ever did was to ensure the quantum mechanicists had their shit together. But starting off your talk by referring to everyone else in your field as "politicians and scientists who want to be politicians" just isn't "god does not play dice" and it sets off every kook alarm I have.
[1] http://www.whitehouse.gov/sites/default/files/microsites/ost...
http://rogerpielkejr.blogspot.com.au/2014/03/john-holdrens-e...
It was extraordinary that Holdren should write such a harsh attack on an individual scientist, particularly when as pointed out by Pielke that his response is entirely consistent with IPCC SREX.
Here is the most important quote in Pielke's article:
What the IPCC SREX (2012) says: “There is medium confidence that since the 1950s some regions of the world have experienced a trend to more intense and longer droughts, in particular in southern Europe and West Africa, but in some regions droughts have become less frequent, less intense, or shorter, for example, in central North America and northwestern Australia.” For the US the CCSP (2008)20 says: “droughts have, for the most part, become shorter, less frequent, and cover a smaller portion of the U. S. over the last century.”21
It appears that stating what the IPCC says is now outside the consensus according to Holdren.
(see e.g. http://publiceditor.blogs.nytimes.com/2013/07/22/nate-silver... )
and now they no longer have any reason to hide it.
Perhaps they could become relevant again by analyzing the actual data as opposed to engaging in this false personality narrative all the time, the kind of personality narrative that elevates corrupt individuals like Chris Christie because, well, he has a big personality. But I doubt they're going to do that.
That said, Krugman is more of a data-minded columnist, as you would expect since he's an economist, so I find his objections to Silver's new site surprising. I don't think Nate Silver ever said that data without interpretation is all there is; but that data with interpretation tells the actual story, as opposed to whatever all these political journalists are doing .
Krugman in his earlier blog posts was more specific:
"I feel bad about picking on a young staffer, but I think this piece on corporate cash hoards — which is the site’s inaugural economic analysis — is a good example. The post tells us that the much-cited $2 trillion corporate cash hoard has been revised down by half a trillion dollars. That’s kind of interesting, I guess, although it’s striking that the post offers neither a link to the data nor a summary table of pre- and post-revision numbers; I’m supposed to know my way around these numbers, and I can’t figure out exactly which series they’re referring to. (Use FRED!)
More to the point, however, what does this downward revision tell us? We’re told that the “whole narrative” is gone; which narrative? Is the notion that profits are high, but investment remains low, no longer borne out by the data? (I’m pretty sure it’s still true.) What is the model that has been refuted?
“Neener neener, people have been citing a number that was wrong” is just not helpful. Tell me something meaningful! Tell me why the data matter!"
krugman.blogs.nytimes.com/2014/03/18/sergeant-friday-was-not-a-fox/
So his objections do take the form of something like you'd expect - this post is just the latest in a series. Krugman clearly had very high hopes for the site - this is his 3rd of 4th post about it, which makes sense given Krugman's prior cheerleading for Silver.
This is one of two or three posts Krugman has made about 538: http://krugman.blogs.nytimes.com/2014/03/18/further-thoughts... http://krugman.blogs.nytimes.com/2014/03/18/sergeant-friday-...
Krugman links to Tyler Cowen saying the same thing: http://marginalrevolution.com/marginalrevolution/2014/03/nat...
And I don't think anyone could make the argument that Cowen and Krugman are backroom buddies.
Krugman has made many, many posts on his blog in full-throated support of Silver and his methods: http://krugman.blogs.nytimes.com/2012/10/28/the-war-on-objec... http://krugman.blogs.nytimes.com/2012/10/25/triumph-of-the-e... http://krugman.blogs.nytimes.com/2012/10/30/scoop-dupes/ http://krugman.blogs.nytimes.com/2009/06/22/the-medical-indu... http://krugman.blogs.nytimes.com/2012/11/03/reporting-that-m...
It's ironic that you're criticizing Krugman on this without looking at the data, which speaks to Krugman very much supporting Silver, and only now expressing distress over the future of Silver's site and his hires (he points out a particularly bad economics analysis by one of the staff and explains his issues with it), not Silver's analysis.
Krugman is a columnist. The NYT cannot tell him what to write (and they have strict editorial practices, so I doubt that they would even try).
Furthermore, a columnist can write essentially whatever they want and the paper will have to publish it. This is part of the agreement when a paper brings on a columnist (and why they must choose them carefully) - they are entering a semi-binding[0] contract to publish anything that the columnist says in the future, without actually knowing what that will be.
[0] I say "semi-binding" because it's not a legal contract, but it is a matter of industry ethics, and the New York Times has a good track record of holding to these.