Saturday, April 9, 2011

More Economic Science/Natural Science Analogizing: Two Thoughts

IS-LM as Ptolemaic Economics?!?!

Here, Mark Thoma paraphrases Larry Summers as saying "Larry Summers just said DSGE models played no role at all in WH policy response to crisis. It was all IS-LM augmented by liquidity trap... Summers says Ptolomy model outperformed Copernicun model for 50 years after discovered. Same for IS-LM vs. DSGE" at the INET Bretton Woods meeting that's going on right now.

Does this really make sense? I don't think so, although it's a turn of phrase that a lot of people are going to love to latch on to and repeat. So what does he mean that the Ptolemaic system worked better than Copernicus at first? Well, Copernicus had some bugs to work out at first - both empirical and theoretical. Empirically, the Copernican system implied that the other planets should have phases, which nobody had observed - leaving many in the Ptolemaic camp. Later, as telescopes were developed that did observe these phases, more people sided with Copernicus. Copernican predictions also weren't even as accurate until Kepler introduced elliptical orbits, and even Kepler's orbits had no solid justification until Newton! But is IS-LM Keynesianism really Ptolemaic relative to DSGE New Keynesianism? I'd say it's more accurate to say that IS-LM plays Copernicus to DSGE's Kepler or Newton. DSGE has a New Keynesian IS curve, after all, and the Taylor Rule and various other monetary policy rules do the same work that the LM curve does as well. IS-LM is still in New Keynesianism, it has just been filled out and modernized. Ptolemy was fundamentally different from Copernicus. Newton didn't throw out Copernicus - he developed him. IS-LM hasn't been thrown out either - the old IS-LM is just a first step toward more complicated models. I personally think in terms of IS-LM because even though I've had graduate macroeconomics and DSGE type models, I haven't had enough of it to really soak in yet. But there's not the fundamental break that Summers implies. I do think Keynes represented a paradigm shift, and I think Classical economics is more analgous to Ptolemy. Nobody except politicians and a few hold-out economists use Classical economics anymore. Even that's probably too harsh, though. Even Classical economics isn't as radically different from IS-LM and AD/AS economics as Ptolemy was from Copernicus. Still, I think there's a good case that Keynes marked a paradigm shift.

In What Way is Economics "Right"?

Here, Phil Plait discusses that kid that has some new ideas about Einstein and likes to do integration by parts on his window. He writes:

"Barnett may very well be a genius, and may very well rewrite a lot of physics… as, no doubt, future generations of genius scientists will. But one thing they won’t do is prove relativity wrong.

Bold statement? Not really. We know relativity is right. It may be incomplete, but it’s not wrong.

What I mean by this isn’t too hard to understand. In science (ideally, if you’ll pardon the pun), an idea becomes a hypothesis, a testable statement. If it passes the test, it can be expanded upon, broadened, tested and retested. Eventually, as it grows and becomes more solid, it becomes a theory — I know, in the general jargon that word means "guess", but to a scientist a theory is an explanation of phenomena so profoundly certain that a layperson would call it a law.

Relativity is just such a theory. It has passed essentially every single test to which it has been put for the past century. It is literally tested millions of times a day in particle accelerators, for example.

So I don’t think anyone, young Jacob Barnett or otherwise, will ever prove relativity to be wrong. What they might do, what I think and hope someone eventually will do, is show how it’s incomplete."

So how does economics stack up? What can we say, like this? Clearly we are different from relativity, but it's not like we're just making things up either. Later on in the post he lists evolution as being with relativity as one of those theories that we know to be "right", and that will never be proven wrong. I think laws in economics are "proven" in much the same way that evolution is (which shouldn't be surprising, since economics is a very, very specialized form of the biology of highly evolved primates). Relativity is "right" insofar as we know precise relations and formulas hold up consistently. Economics isn't "right" in that way. We don't have a supply curve that is the "right" supply curve. It sounds strange even to talk about it in those terms.

What we have, like evolutionary biology, is a mechanism that we know functions in the way we think it functions, broadly speaking. We know without a doubt that organisms evolved by natural selection, just as we know without a doubt that the price mechanism optimizes resource allocation (with "optimize" having a specific definition in this case, just as "selection" has a very specific definition in biology). We know the process is incontrovertibly true, but the way that process plays itself out is very contextual and historically contingent. When did humans evolve certain features or move to certain regions or emerge as a distinct species? These are natural history sorts of questions that are informed by the theory of evolution. These specific questions are analagous to questions like "how elastic was labor demand a decade ago?" or "what drove the inflation rate in the 1970s". The scientific law of supply and demand and market efficiency informs how we answer these questions both theoretically and empirically, but there is a lot more natural history to it than there is with relativity. Astrophysics has its own natural history, of course. People that try to answer questions about the origin of the moon or the asteroid belt set themselves to answer questions that are just as messy as the ones that biologists or economists deal with on a more regular basis.

Scientific theories are "right" in a lot of different ways, and people that are down on economics as a science should put aside their physics envy and just recognize we aren't "right" like relativity is "right". We're more "right" in the way that evolution is "right". We know that we have an accurate understanding of a broad mechanism or process that is central to answering the questions of economics, just like evolution is a broad mechanism or process that is central to answering all sorts of questions in biology.


  1. "Scientific theories are "right" in a lot of different ways, and people that are down on economics as a science should put aside their physics envy and just recognize we aren't "right" like relativity is "right". We're more "right" in the way that evolution is "right". We know that we have an accurate understanding of a broad mechanism or process that is central to answering the questions of economics, just like evolution is a broad mechanism or process that is central to answering all sorts of questions in biology."

    How about we say that ideas are, just like products or services of businesses, a discovery procedure? Like what Hayek called it?

    Nobody thought a Sony Walkman would bring in anybody's interest, being a crude, malfunctioning device with bad sound. However, we didn't know how the product would be accepted until it was actually tried.

    The same way, we can't know in advance what questions evolution will answer until those questions actually come up, and we can't know what phenomena all the past economics theories will explain until a situation actually comes up that could be used for it.

    For all we know, they could sit on the backburner until some strange event justifies a long discarded idea from medieval era economics.

  2. "These estimates, like the aggregate ones, are subject to substantial margins of error. One additional source of
    uncertainty concerns the impact of the state fiscal relief. We believe that the rule of thumb that 60% of funds devoted
    to state relief will be used to prevent spending cuts and that 30% will be used to prevent tax increases, and that these
    effects will occur with a lag of about three months, are good first approximations."

    "And federal aid to state and local governments wasn’t enough to make up for plunging tax receipts in the face of the economic slump. So states and cities, which can’t run large deficits, were forced into drastic spending cuts, more than offsetting the modest increase at the federal level."

    Romer, lead economic scientist (snort) made a bunch of overly-rosy claims about the stimulus while also saying that the margins for error were so large that her estimates were highly suspect (Science!) She accounts for the way that states will respond to the stimulus but (according to Krugman) somehow missed the fact that there would be lower tax receipts!? How is this possible? How can the most obvious of issues not be incorporated into these super-important super-scientific calculations? If they are, why do they produce answers so far from reality? Why can Krugman say with a straight face that Romer didn't account for lower tax receipts in her calculations? If macro is a science, why is it that only Krugman can get it right... and why is his fellow Nobelist not employing him instead of all these knuckleheads who can't remember that whole state tax thing?


All anonymous comments will be deleted. Consistent pseudonyms are fine.