Comment by klochner | Hacker News Reader

Comment by klochner | Hacker News Reader

Am I the only one who read Friedman's paper? [1] The answer to OP's question [2] - "why are economists unaware of milton friedman's thermostat" - is because the OP is misinterpreting Friedman's paper, and therefore has misattributed this nonexistent "theory".

Friedman just said that the Fed acts like a thermostat, and in the linked paper, claimed that the period subsequent to 1985 had increased price stability owing to improvements in their understanding of inflation and it's relation to the quantity of money, hence their "better thermostat".

AFAICT, the OP has made up this "Friedman's Thermostat" insofar as it relates to economic forecasting, and now he wants to know why no one has heard of it.

As an aside, "exception that proves the rule" is an almost universally misapplied phrase, that should only be used if you're really really sure you know what it means (kind of like "begs the question").

[1] http://online.wsj.com/article/0,,SB106125694925954100,00.htm...

[2] OP = Nick Rowe - http://worthwhile.typepad.com/worthwhile_canadian_initi/abou...

Am I the only one who read the blog article? It's an old idea that shows up in the literature of various smart economists, and is in fact just a name for a common phenomenon in statistics, with the twist that human behavior causes it. He's not criticizing people for not having heard of the name (which he seems to have made up in 2010), he's criticizing people for being unaware of the idea despite it being an old idea and, he alleges, a very important one when you start studying human behavior. All of this is in the blog post.

cynicalkane: "He's not criticizing people for not having heard of the name"

Nick Rowe: "Google seems to tell me I'm right. I'm the first link, which is really pathetic for such an important idea."

Not to mention that googling for "Fed Thermostat" brings up a lot of references. Friedman's is in the top page, Nick's is not:

https://www.google.com/search?q=fed+thermostat

Well, people being unaware of the name doesn't seem to have anything to do with the rest of his thesis. So if you're correct that would be rather incongruous with the rest of the post, which I think is valuable and worth reading.

So I don't think he's criticizing people for not knowing the specific name, but there's some egotism at play. He likes the idea he identified and the name he came up with, and is scratching his head that only market monetarists have picked up the name and that his blog--minor in the blogosphere--is still the first link.

Full quote:

Google seems to tell me I'm right. I'm the first link, which is really pathetic for such an important idea; the second is Friedman himself (pdf); and most of the rest on the first page are other bloggers, mostly Market Monetarists. But this idea has got nothing (in particular) to do with Monetarism.

But arguing about the interpretation of this sentence ultimately is a red herring--you don't go into rage-mode about a blog post because most of it says X but one sentence seems to say Y instead, especially if you can't be sure it even says Y.

The main thrust of his post is "look at this idea that no economists seem to be aware of". He makes his point by googling for this term that he made up.

He does not appear to be an Econometrician, nor a Statistician (I can't find his CV, which is odd), so the most likely answer is that he is wrong, and that this is a well understood phenomenon for which he doesn't know the proper technical term (viz., Multicollinearity).

This quote from his blog post sums it up for me (emphasis mine):

"And it really bugs me that people who know a lot more econometrics than I do think that you can get around the problem this way, when you can't"

[edit] - @cynicalkane - we are in violent agreement on the core theory, but equally violent disagreement on the extent to which his link-baity title & appropriation of Friedman's name is academically disingenuous.

Shall we call it a day?

That's a strange thing to complain about, because it's accepted that outsider criticism of econometrics is valid, including among econometricians. It's common to question the assumptions even if you don't understand the machinery. There's been a lot of it in particular about market modeling in the face of the subprime mortgage collapse, both from economists and non-economists.

I'm also not sure why you keep blasting him for not using the word "multicollinearity", when he's talking about a fairly specific kind of hidden variable anyway, and one that you simply can't pluck out of small macro data sets ex nihilo especially in the face of the Lucas critique.

For instance, in this blog post, http://ipeatunc.blogspot.com/2012/08/its-not-just-data-its-a... , a blogger discusses Milton Friedman's thermostat and then, in the comments, discusses deliberately leaving technical terms such as multicollinearity out of the post.

"Are you aware of this idea? Maybe under some other name??"

First paragraph of blog article.

Right, if you find yourself asking "Why has nobody else noticed this" you should always consider that the answer is "because it isn't there." But on the other hand, the idea that intentional behavior can remove correlation from causal relationships is important and deserves more thought than people give it. And since economics has something of a history of naming things for the wrong person, calling this "Friedman's Thermostat" and spreading it around might be a decent idea after all.

The delivery is so emotionally charged too. It took several paragraphs of "I can't believe no one gets this!" before we even got to Milton's theory.

It's worse than that. He rambles on, and then immediately shifts the metaphor to a more complicated and less clear one. He should have just stuck with the thermostat: if the temperature inside a house is determined by a thermostat, the outside temperature has no correlation with the inside temperature.

I suspect this is because he's already written several other posts (linked at the bottom) with the thermostat example, so is trying a different example here, in the hope that people might pay more attention this time.

Agreed. An analogy made by Friedman in a WSJ op-ed != established theory.

Am I the only one who read the Casey Mulligan blog post which kicked off Professor Rowe's rant? Actually I didn't read the whole thing; I couldn't force myself to push past the egregious statement that "the Federal Reserve’s monetary policy has little effect on a number of financial markets, let alone the wider economy." Anyone who lived through the recessions of the early '80s in the US and believes that is sadly unable to learn from experience.

Sorry, what do you mean "OP is misinterpreting Friedman's paper"? I don't think there's much "interpretation" here. He's basically saying: here is a paper about X. It is titled "X".

And I'm also confused by what you mean by "the OP has made up this 'Friedman's Thermostat'"... seriously, what does that mean? Friedman wrote a paper on Friedman's Thermostat. So how could OP [whoever that is, I assume you mean Nick Rowe] have "made it up"?

the article is "the FED's thermostat" as an analogy about the Fed's. There is no "Thermostat Theory" it's just a title, not a general theory.

Huh? Yes, "thermostat" is used as an analogy (by both Friedman and Rowe).

But the article is not claiming that there is a "general thermostat theory." He's talking about one idea (there may be no measured correlation under targeting) and asking why this is not more widely known, especially since the abstract idea seems more widely applicable (to fields other than monetary theory and applied econometrics).

The inverse of Rowe's claim is that this idea is well known (maybe under a different name). Are you arguing that? If so, then by what name?

He's saying two things: 1) I am the only one who knows about this terminology I invented? ("maybe under some other name?", well, how could people know for this name? You invented on your blog, for christ's sake. Obviously you are the first result on Google. Egomaniac.) 2) Then he talks about this abstract Idea, which he thinks is not well known. About the first topic. WTF. About the second topic, not sure if it's such an ignored aspect, but ok, it's a valid point.

OK, I see where you're coming from. I don't think he's saying 1). But then again, I've been reading him for a while, so I had a bit of context. If this had been a random one-time post on some guy's blog, then sure, I'd probably think 1).

>Friedman wrote a paper on Friedman's Thermostat.

No he didn't.

He just used the WORD "thermostat" once, in an article.

And his use of the word was unrelated to the concept presented in the blog post (which is not Friedman's but older and which nobody calls "Friedman's thermostat" except the author of the blog post).

Here's an analogy for the situation:

David Pogue writes and article titled "The perfect gadget for Christmas", which is simply a review of 10 hot current gadgets.

Then, 30 years later, some blogger writes "How come nobody knows "Pogue's Perfect Gadget" theory? You know the one that says that when you put a cat inside a gadget's enclosure with a quantum based poisoning device the cat's state is both dead and alive?"

He then goes to admit that idea: is not Pogues, it was found earlier, and nobody calls it that -- all the while complaining why people don't know it by his made up name "whereas they do some other Pogue's articles".

The whole blog post smells crank to me...

Let's get the obvious out of the way: The thermostat concept is an analogy. In Friedman's original article, it's being used to describe money supply targeting [not real thermostats].

Nick Rowe is abstracting away some of the specifics of Friedman's article, but he's talking about the same concept : how active targeting of one measure can achieve desirable changes in another one while still showing no/little correlation between the two.

Rowe is also applying that abstraction to another concrete example: the pedals in a car driving in a hilly area.

Reply on news.ycombinator.com