Wednesday, January 13, 2016

More on scope

So Simon Wren-Lewis teaches me something I didn't really know:
The big models/schools of thought are not right or wrong, they are just more or less applicable to different situations. You need New Keynesian models in recessions, but Real Business Cycle models may describe some inflation free booms. You need Minsky in a financial crisis, and in order to prevent the next one. As Dani Rodrik says, there are many models, and the key questions are about their applicability.
Ok, I knew about what Rodrik was saying, but I thought that applied to situations like liquidity trap versus away from the ZLB. I was a bit shocked to learn that it applied to Keynesianism versus RBC (at least according to Simon Wren-Lewis).

However nothing in those models says anything about their limited scope. RBC DSGE models do not say they apply under conditions X (during inflation-free booms), and new Keynesian DSGE models do not say they apply under conditions Y (during recessions). Both theories lack scope conditions (or at least they never seem to be stated).

Minsky's model for example explicitly says it operates across the entire business cycle. You can't just relegate it to financial crises -- essential to the model is about how finance builds in a boom-bust cycle. Minsky applies during recessions at the same time as NK DSGE does and during inflation-free booms like those RBC DSGE models.

I do think this helps us understand a bit about macro, though. Macro appears to be made up of several pieces of the elephant, but without a framework and scope conditions we have no idea how these pieces fit together. I think it is related to how economic theory doesn't really address how one travels from one equilibrium to another***.

Which means there is a really good analogy in physics: string theory. In string theory, there are five different theories (pictured at the top of the post) -- each thought to always apply (infinite scope) since they were ostensibly fundamental theories of everything. These individual theories are speculated to be just different pieces of an elephant called M-theory (that has a low-energy limit called 11-d supergravity).

Economics hasn't really discovered its S- and T- dualities yet (the things that relate the five string theories to each other), nor does it have the common framework (quantum field/string theory) for the different theories ... so it probably has a ways to go.


Update + 1 hr

25 minutes later, Noah has chosen the same quote for his blog post. I kind of hint at his #3, but mostly discuss his #2 above. His #1 is something that didn't occur to me as a scientist: I would never consider figuring out what model/scope is appropriate only after the scenario happens. Sort of a "Keynes works during recessions, therefore Keynes is the appropriate model for recessions" conclusion. But it is very important!

Update, the second

I am pulling out the quote from David Glasner I referred to above (in the sentence with ***):
But there is no theory describing the laws of motion leading from one equilibrium to another, so the whole exercise is built on the mere assumption that a general equilibrium is sufficiently stable so that the old and the new equilibria can be usefully compared. In other words, microeconomics is predicated on macroeconomic foundations, i.e., the stability of a general equilibrium. The methodological demand for microfoundations for macroeconomics is thus a massive and transparent exercise in question begging.

... and juxtaposing it with a quote from Noah's piece above (in the first update)
But in general, economic models are hobbled by the fact that the underlying theory itself (the "microfoundation") is in dispute. Physics models are usually just approximations of a single underlying theory. Econ models actually contain different assumptions about the nature of human and institutional behavior.

One: my speculative sentence marked with *** above was based on some gut feelings, but those gut feelings are leading to some warm fuzzy feelings after Noah's quote.

Two: Physics models are usually just approximations of a single underlying theory. That's why I put up the string theory analogy: it's a case where the underlying theory isn't known.


  1. I am surprised to see the phrase "Scope conditions" appear on a macro-themed blog other than yours (i.e. Noah's). Did that surprise you?

    Not being familiar with M-theory (beyond what I can recall from a Brian Greene youtube video aimed at people like me) or any of its five satellite theories (in your diagram), I'm slightly confused by your point you make in Noah's comment section here:

    ""Physics models are usually just approximations of a single underlying theory."

    I think that is a good point; it also means there is an additional example from physics that is more like the econ situation: string theory."

    So, looking at the context of Noah's original quote there, he seems to be saying "Macro has a problem physics doesn't have" and you're saying "Good point Noah, but actually here's an example from physics that demonstrates an analogous problem to what you're bemoaning about with macro." Am I at least getting the tone right there? I.e. to put it in caveman (the voice in my head):

    Noah: Macro bad. Physics good.

    Jason: Yes, but here example where physics bad in same way too.

    1. I got scope conditions from Noah. I would say scale of the theory or domain (of validity). (Click scope link above for reference.)

      I wouldn't characterize it as "bad". It's the state you find yourself in when you don't know what you're doing.

      Economics is where string theory was in the 1980s before the idea of M-theory. Strings (microfoundations) didn't lead directly to enlightenment, so there was confusion and competing theories with no idea of their scope.

      Witten comes along with M-theory and what people think is a path to enlightenment.

      Maybe, maybe not.

    2. Scope: I guess I'd forgotten where that come from.

      So M-theory is a theory which "isn't known?"

    3. M-theory is a conjecture -- it's conjectured properties include 10-dimensional holographic representations (on the 10-d boundary of an 11-d space) as each of the 5 known string theories and a low energy limit as 11-dimensional supergravity (supersymmetric general relativity).

      Witten has said M stands for 'mystery', but some also think M is just and upside-down W.

  2. Maybe economics is where medicine was before the germ theory of disease.

    1. That's occurred to me too. But sometimes I wonder if economics is especially hard because it's a moving target: perhaps it changes so rapidly relative to the data it produces so as to make it a tough natural phenomena to track. I love that Jason takes a crack at it though.

      One of the reasons I'm such a fan of this blog is the good mix of theory, philosophy, comparisons with data, and forecasts. Noah Smith is good on the philosophy. Nick Rowe is all theory. Lots of people attempt to explain data, but rarely do you see a curve showing what their (well defined mathematical) model says in comparison. John Cochrane is a good example: he shows plots about what his models do, and he shows plots of actual data: but I don't recall him mixing the two. And very few macro-bloggers show plots of what their models forecast (followed by subsequent updates about how well those forecasts did). I wish there were more blogs that had this same mix.

    2. Cheers, Tom.

      Bill, the only issue with that analogy is that we do kind of know what the "agents" are: people making transactions using money. People couldn't see germs until microscopes were invented.

      But a big point I try to make on this blog is that maybe being the same scale as the economic agents (and being agents ourselves) has blinded us to other ways the system could work. Maybe our actions aren't everything and the state space is what is really important?

      It would be like if doctors were at the same scale as bacteria ... and that blinding them to the idea that genetics or environmental factors (e.g. carcinogens) can cause things because they think it's all just bacteria ...

      An analogy I made early on this blog was that I wanted to look at the economic system on Earth as if I was an alien on another planet seeing the exponential burst of CO2 through a telescope spectrometer -- and positing a "civilization" operating under a system of "economics" as a model.

    3. What I had in mind was what Noah says about expert judgement. Before the germ theory, medicine may have been less effective that it is today, but it was not completely ineffective. But the practice of medicine relied greatly upon expert judgement. (It still does, but not so much. Why make a diagnosis that has 2:1 odds of being right when you can run tests?)

      BTW, I, too, was taken aback by Noah's point #1. Yes, post mortems can have value, but they can also be opportunities for face-saving and begging the question.

      There is a story about a psychiatrist who had a patient who believed that he (the patient) was dead. At one point the psychiatrist pricked the patient's finger with a needle, drawing blood. The patient looked at the drop of blood on his finger, experiencing apparent mental turmoil. Finally he relaxed and said, "That just proves that dead men bleed." ;)

    4. Thanks, Tom. :)

  3. O/T: Were you aware of this apparent near convergence of Farmer/Sumner/(Rowe?) thinking?

    1. Not sure the convergence is terribly deep ... in any case, I am trying to write down Farmer's model

      PY/W = b L

      as an information equilibrium relationship. Actually, for some reason I almost did write that down on 2 November 2015 (it's just missing the b)

      But the reason escapes me. It might have been in relation to this post:

      but I honestly don't remember.


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.