Wednesday, May 06, 2009

Macro with Incentives?

Tyler Cowen and Alex Tabarrok have a new Macro textbook, and it sounds awesome. It actually prioritizes incentives. Incentives-based approaches basically apply the assumption of individual self-interest and looks at the implications. That's Adam Smith's insight, and he's the father of economics. The father of macroeconomics has not generated similarly fruitful insights. Since Keynes macro has focused on aggregate accounts and indices, operating according to laws of motion. Sure, you might motivate the laws via a consumer maximizing his lifetime income, but that's all throat clearing for estimating equations where aggregate amounts C, G, I, and X interact with monetary variables P, i, and M in some dynamic feedback loop

The hope was that macroeconomics would be like physics, where you have laws at a lower level, and unrelated laws at a higher level. In 1972, Philip Anderson, who won the Nobel prize in physics for his work on superconductivity, wrote an article titled “More Is Different,” and contended that particle physics, and indeed all reductionist approaches, have limited ability to explain the world. Reality has a hierarchical structure, but that does not mean one should always try to explain one layer from deeper layers.

At each stage entirely new laws, concepts, and generalizations are necessary, requiring inspiration and creativity to just as great a degree as in the previous one. Psychology is not applied biology, nor is biology applied chemistry.

No collective organizational phenomenon, such as crystallization and magnetism, has ever been deduced from its lower-level parts. Emergent phenomenona render reductionist views irrelevant for explaining phenomena at that level—it is unhelpful to try to understand cancer through mere chemistry, or worse, particle physics. But that does not mean that chemistry, or particle physics, is uninteresting, just, at one level, not so much.

Unfortunately, macro economics has not generated comparable emergent properties in macroeconomic time series. Incentive matter a great deal on the macro level, even if they are more difficult to reify than something like total consumption or the money supply. The stagnation of socialist economies was not predicted by 1950's economists, indeed most economist thought capitalism was merely superior in liberty, not growth or prosperity. The thought the economy was like a big input-output matrix, and with centralized planning combined with forced savings, the Soviet Economy seemed sure to overtake the West. But this turned out quite wrong because people in a position to make decisions did not have the right incentives. I don't know exactly how Cowen and Tabarrok handle this in their text, but a good assumption goes a long way, and moving from textbooks with price-stickiness models, or government multipliers, to starting with incentives, is a first-order step in a better direction.

They mentioned they address the current financial crisis, and though I'm sure there's a demand, with so much new information coming out, and the fate--if not the current condition--of banks so uncertain, I sense that section will be much rewritten if there is a second edition. In any case, mazel tov, gentlemen!


Michael F. Martin said...

Woefully uninformed post. Crystallization and magnetism have been explained to a good approximation by renormalization group, the ising model, kadanoff transforms, and a host of other concepts from non-equilibrium statistical mechanics. Just because economists haven't been keeping up with physics doesn't mean it hasn't been done. Sure thermodynamics and stat mech are different theories with different scales in which they apply well as approximations. But just like micro and macro, they're describing the same reality, and should ultimately be reconcilable. If not, then one must be incorrect.

Eric Falkenstein said...

That's not what Nobel prize winning physicist Robert Laughlin thinks. He wrote a book on just these matters in A Different Universe. I don't think the ising model is based on the four forces in any essential way, it is just a way to model interacting forces, but that's not really the same thing as saying Bohr's model of the atom, or Dirac's model of the electron, implies chrystalllization.

As to reconcilability, that's irrelevant. Potentially gravity is reconcilable with strong forces, but I don't see how any such reconciliation would be fruitful empirically (yes, it's pretty). You can make any macro theory compatible with any micro theory with enough creativity, the bottom line is that the macro equations applied to aggregate variables don't work.

Michael F. Martin said...

Maybe I overstated things. But hey, it's a blog comment. Favorite soap box for windbags, right?

Here's what you wrote in your post:

No collective organizational phenomenon, such as crystallization and magnetism, has ever been deduced from its lower-level parts. My point is that any of the theories I mentioned provide a reasonable explanation for how the macroscale phenomena of phase transitions is built up from microscale dynamics.

I haven't read Laughlin's book so can't make a specific response there. I do know that Laughlin is coming at his point of view from having spent a career studying plasmas, which are a great deal more complex to model than the average crystal or magnet. The economic analogy might be to the difference between modeling the cotton market and modeling the oil market.

Unknown said...

Somewhat off topic, but there was a recent conference on economics at the Perimeter Institute with actual theoretical physicists, and Derman quoted one of your blog posts in his talk. Video here:

Eric Falkenstein said...

that's fun :)