Thursday, June 2, 2016

Economics is wrong, therefore (by analogy) physics is wrong

Since this mentions economics, it lets me talk about it in a way that is germane to this blog.

Let me get this out of the way: anyone who thinks global warming doesn't exist, global warming won't warm the planet by a few to several degrees over the next hundred years, and/or global warming won't be disastrous for Earth is basically a terminal fool. Whatever the reason -- Dunning-Kruger, bias (money or tribal), idiocy, or terminal contrarianism -- there is one; you just have to find it. So let's go!

Megan McArdle makes this seriously flawed analogy:
Economists can't run experiments in which they change one variable at a time. Indeed, they don't even know what all the variables are. ... Because climate scientists, like the macroeconomists, can’t run experiments where they test one variable at a time, predictions of feedback effects involve a lot of theory and guesswork.
McArdle has an English degree and an MBA; if this isn't Dunning-Kruger, I don't know what is. 

False equivalence 1:

In one case, the fundamental theoretical principles of the micro theory are known and tested to multiple decimal places; and in the other there are completely false assumptions of rational utility maximizing behavior and Calvo pricing.

Knowing the micro theory allows you to theoretically isolate variables to study the macro theory. Therefore there is no requirement about experiments isolating one variable at a time. If there was, astronomy would be in some pretty serious trouble because it can't isolate any variables.

In my personal opinion, this claim rises to the level of requiring a retraction of the article. It's like claiming 2 + 2 = 5; it's completely wrong.

False equivalence 2:

In physics (and the natural sciences in general), there is a principle that we call "naturalness" where we use scales to understand the system. The choices for your dimensionless coefficients (numbers without units, having been removed by identifying scales) in your theory are restricted to the set of 0, 1, or infinity. You can think of infinity as the "dual" of zero that tells you that one-over your dimensionless number is zero. Choosing some other number -- like a thousand or a ten-thousandth -- requires an explanation. That's at the heart of the hierarchy problem as well as the strong CP problem in physics (they're sometimes called fine-tuning problems). A few days ago I responded to a commenter using exactly this naturalness principle with respect to global warming:
The Earth converts a few high energy photons into many low energy ones (entropy production). Due to that conversion, the existence of CO2 absorption in the infrared, and the random emission direction of the absorbed photon (quantum mechanics), the final result is a higher temperature on Earth with CO2 than without. Without any CO2, the Earth would be about 40 K colder (the temperature of the Earth without atmosphere). Obviously changing CO2 levels by upwards of 20% should -- assuming the coefficient is natural -- produce approximately 20% of this warming effect. That would be about 8 K -- or at least its equivalent in thermal power divvied up among the water and land, which is the the more uncertain bit. Even a silly allocation where 75% of the power goes into the water and doesn't raise its temperature at all leaves you with 2K. Both the 8K and the 2K are perfectly in line with the range of climate forecasts. 
Any of the arguments against global warming have to claim an unnatural coefficient for the impact of CO2 (i.e. a 20% change in CO2 leads to x << 20% of temperature change), which is why they'd tend to be dismissed out of hand -- at least by physicists. It's just not consistent with what we know about CO2. And if they claim an unnatural coefficient for CO2, then they have to explain why the Earth is 300K and not 250K.
This naturalness principle allows you estimate the effects of perturbations to variables without necessarily controlling the other ones.

In traditional economics [1] there is no such principle because coefficients could have been chosen by human decisions, bureaucratic inertia, or through some strange socio-political process (think 2% inflation). You can't make estimates of theoretical effects in economics because you can't isolate scales (almost everything is dimensionless except time) and use naturalness arguments.

That's basically it. Unless it contains an argument against the naturalness of the impact of CO2 absorption in the infrared, anyone's skeptical position can be dismissed out of hand. That goes for "lukewarmists" as well. Global warming predictions make sense given the back of the envelope calculation I did above, and it would take something comparable to convince any scientist otherwise.

McArdle's argument is that because economics has no scales, naturalness, or fundamental theory, physics (by analogy) has no scales, naturalness, or fundamental theory. Using that argument you could say that because economics is not empirically successful, physics (by analogy) is not empirically successful. I hope the sheer idiocy of making such a stupid argument is clear.



Someone linked to this post in comments on a pro-free market blog, which prompted a response (quoting me above and adding emphasis) that gives an excellent counterexample of the proper use of the null hypothesis:
Obviously changing CO2 levels by upwards of 20% should — assuming the coefficient is natural — produce approximately 20% of this warming effect. 
The assumption is invalid.
If the assumption is invalid, there must be a reason. That is the entire point of the strong CP problem (the assumption of a natural coefficient for the CP violating term in the QCD Lagrangian is orders of magnitude off, for which the axion is offered to explain). And if the assumption is prima facie invalid, its negation should be a valid starting point -- i.e. assume the coefficient is unnatural. Sounds rational! And if the coefficient is unnatural, why is the earth not a ball of ice?

This commenter would look at the hierarchy problem and say: What's the big deal?



[1] In information equilibrium, naturalness should hold. However for non-ideal information transfer all bets are off.


  1. A response to that response to the comment (by commenter "Neo") that you mention in your Update reads as follows:

    The best analogy for CO2 heating is to think of a glass box containing bees. What are the chances of hitting a bee when there are X in the box. What is the probability when it’s 2X .. then 3 X, etc.
    Eventually, you see that bees (and CO2 molecules) aren’t points, they have size, so at some point, the more bees you added .. there is no change.
    This is why the relationship of CO2 concentration to absorbing sunlight is asymptotic .. and we’re already above the knee of the curve.

    I have no idea what that means. Do you?

    1. Hahahahaha!

      That is ridiculous.

      Regardless of the density, we're still talking about CO2 gas. When CO2 molecule size matters, you must be close to the solid-gas transition (CO2 sumblimates at 1 atm or less). But then even then the thickness of the dry ice would matter for absorption. Having a 100 mile thick slab of dry ice would probably absorb all of the infrared photons and that commenter says it doesn't.

      CO2 also isn't absorbing sunlight (5000 K) so much as absorbing the infrared light radiating from the Earth because it is 300 K.

    2. OK, that helps... so you think he's saying that the CO2 molecules are packed in so densely, they're like those multicolored balls they put in kids' bouncy houses? And since we're already there, then adding more can't cause any more damage? OK, if that's what he's saying, that is funny.

  2. Megan McArdle is just wrong most of the time. As a pseudo-intellectual libertarian-lite defender of the status quo and opponent of anything to reduce inequality, she is assured steady employment as a useful villager idiot. The fact that she is wrong about just about everything never seems to bother her employers.

  3. Jason, a bit off topic, but I listened to his yesterday and thought it was interesting:

    There's a transcript as well. One of the things the host (Julia Galef) and the guest (David McRainey) briefly touch on is the idea that people carry models (sometimes very crude or inaccurate) models around with them in their heads about how the world works, and it's rarely practical or advisable to expect or to attempt to change someone's mind about something by thinking the evidence you present will leave them with no other option but to completely discard their old model of the world. They make an analogy between this and how "science" changes its "mind." I don't think either of these two people could be called a scientist though.

    In terms of the average non-scientist Joe off the street, getting them to change their mind with evidence alone can be exceedingly difficult (David argues). As an example of a case where evidence did work, he brought up a well documented case of a prominent English 9/11 Truther, who's mind was changed after we was taken on a tour of evidence: including talking to structural engineers, demolition experts and even the mother of one of the victims. And still, only 1 of 5 Truthers given this same treatment changed their mind. They compare that with a different technique of getting people to assign a level of certainty to their beliefs, and then ask about how they arrived at that figure. This combined with non-judgemental and interested attitude about what the subject is saying can with some success (~10%) get people to reexamine their ideas and question where they obtained them, and question if those are really their ideas. This last bit perhaps has little to do with science, but if you're dealing with people not used to thinking in a scientific way, I wonder if there's something to be said for looking at more successful ways of inducing a reexamination of views. I.e. I wonder if this is at all useful in the world of economics. =)

    Some excerpts from the transcript:

    If it wasn't that kind of appeal, it would be just being a citizen of the internet, talking to people on Facebook or Twitter or especially science communicators who are people who are in the hard sciences ... I'm going to be slaughtered by my social science friends, but people who are in the hard sciences often come up against this when they're actually dealing with human beings and they try to say, "Look, this is my life. This is what I do for a living. This is what I've done for the last 30 years. I've been to Antarctica. I've got core samples. I have the facts. I have them. I have the mountain of them and they're all on my side. Why won't you listen to me?"


    It's amazing the parallels between the way science changes its mind and the way if an individual is going to change his or her mind, it follows a very similar path, which is the whole thing doesn't go at once. Some anomaly here or there starts to make it so that your current model of reality isn't making as good of predictions as it should be making. It isn't leading to behaviors and interpretations that are giving you the kind of results that the model was giving you in a previous sociological environment. It necessitates that you update it and change it in some way so that you can continue to make good predictions and you can continue to put things into proper context and it's very similar.

    1. It sounds like the success rate of the evidence only method (at ~20%) is better than the other approach (I'll call it the Socratic Method for lack of another term) at 10%, but the point was the Truthers weren't just emailed a link to look at... they took them on a tour and introduced them to people in person who knew a thing or two about the subject. Clearly WAY more evidence and attention from experts than most Truthers will ever encounter (in a visceral, concrete in your face sort of way).

      BTW, the Truther's name is Charlie Veitch. Once he "left the cult" his former fellow Truthers set out to destroy him. Here's a bit on him:

      Of course I'm making the brash assumption that you are not a 9/11 Truther... =)

    2. Generally, I've proceeded in life using Planck's dictum about science progressing one funeral at a time. The arguments between X and Y don't change X or Y's minds. What they do is allow a younger generation Z to either go along with X or Y. Usually -- at least one hopes in science -- the better arguments of Y attract more Z's, so progress is made.

      That's the 80% model. That other 20% of X's who change over to Y helps a bit too.

    3. You're probably right. Although McRainey says that's what he USED to believe before writing his book (essentially Planck's dictum) but doing the research for his book changed his mind and now he's a bit more hopeful, although maybe he doesn't mean by that any more than that extra 10% (maybe he thought that was closer to 0% prior). I might have to buy the book to find out.

      Here's a disturbing thought: if we are able to significantly extend life expectancies one day, does that imply science slows down as a consequence?

    4. I guess it strongly depends on what that 10% means. If it is a 10% chance each time one is confronted with evidence, then it takes a little over 20 confrontations before someone is converted 90% of the time.

      If the 10% represents a type of person who can be converted, then while that is a bit more optimistic than 0%, it still says 90% of people don't weight new evidence with regard to claims.

      And yes, that would mean life expectancy would impact the pace of science negatively, ceteris paribus. However the network effect of additional people in science as the population grows would probably swamp that effect.


Comments are welcome. Please see the Moderation and comment policy.

Also, try to avoid the use of dollar signs as they interfere with my setup of mathjax. I left it set up that way because I think this is funny for an economics blog. You can use € or £ instead.

Note: Only a member of this blog may post a comment.