I do not think that the currently popular DSGE models pass the smell test. They take it for granted that the whole economy can be thought about as if it were a single, consistent person or dynasty carrying out a rationally designed, long-term plan, occasionally disturbed by unexpected shocks, but adapting to them in a rational, consistent way. I do not think that this picture passes the smell test. The protagonists of this idea make a claim to respectability by asserting that it is founded on what we know about microeconomic behavior, but I think that this claim is generally phony. The advocates no doubt believe what they say, but they seem to have stopped sniffing or to have lost their sense of smell altogether.
…The DSGE school populates its simplified economy – remember that all economics is about simplified economies just as biology is about simplified cells – with exactly one single combination worker-owner-consumer-everything-else who plans ahead carefully and lives forever. One important consequence of this “representative agent” assumption is that there are no conflicts of interest, no incompatible expectations, no deceptions.
And no Recalculation problem, in that the single representative agent never has to re-arrange patterns of specialization and trade.
Pointer from Simon Johnson. On this issue, I am still a Solow student. I share his dismay at the path that macroeconomic theory took over the past thirty years.
Paul Krugman shares this dismay also. He is ready to go back to the old-time Keynesian religion. I am trying to be completely agnostic, taking into account the history of macroeconomic controversies but ready to proceed from a blank sheet of paper. Solow, in his written remarks, does not reveal where he stands. My guess is that it is somewhere in between Krugman and me, with policy leanings closer to Krugman’s.
READER COMMENTS
Doc Merlin
Jul 22 2010 at 8:56am
How about this for a start?
What we need are multiple agent based models. Models with different not identical representative agents.
Its very similar to what Lin, Ren and Sornette did in their finance bubble paper.
“A Consistent Model of ‘Explosive’ Financial Bubbles With Mean-Reversing Residuals.”
Charlie
Jul 22 2010 at 9:40am
How is the recalculation different than DSGE models with sectoral shocks and incomplete mobility across sectors?
John DGSE
Jul 22 2010 at 11:48am
This post seems a little inaccurate, as does Solow’s testimony. One can add wedges (labor, tax, etc.) into a Dynamic Stochastic General Equilibrium model easily, and it is done on a regular basis.
These wedges can easily stand in for such “conflicts of interest”. Take an efficiency wages issue, turn it into a labor wedge, and DSGE solves your issues.
Rebecca Burlingame
Jul 22 2010 at 11:48am
There is an interesting parallel that indicates the needed link between both macro and micro:
Think of the individual who – despite his best efforts and decades of careful decision making – has not managed to create a secure future in any way imaginable. Not just true for the individual…it is also quite true for governments.
A multiple agent model (if this non-economist understands Doc Merlin at all) would be to find a way to link the human equivalent with actual creation of wealth, one not measured by the same criteria as fiat money.
I believe that fiat money and the human model intersect but actually work along different resource use models, and that the real task is to find ways for them to intersect and reinforce one another.
Curt Doolittle
Jul 22 2010 at 11:59am
“How is the recalculation different than DSGE models with sectoral shocks and incomplete mobility across sectors”
In terms of a social science research program:
1) Granularity, and its emphasis on avoiding error-of-aggregation.
2) Granularity, and its greater ability to identify possible shocks.
3) The implication that we should often bypass the financial system to correct shocks.
4) The resulting policy implications that would a) increase relevant information, b) expose our necessary ignorance, c) identify new points of research rather than our current emphasis on data mining and aggregation errors, d) use this knowledge to focus the state on its purpose: competitive investment by the pooling of resources rather than it’s emphasis on exploiting the economy for purely redistributive ends.
Aggregation emphasizes monetary policy. It manufactures political ignorance, and implies gaussian stability, rather than the fragility that represents the actual economy. Shocks are not rare, they are the norm. And the shocks you do not see coming are simply a result of the lack of granularity of your model.
We need to invert economics, so that it does not seek the normative (the normative is an illusion of the process of human memory biases), but build economic research that seeks the shocks, at all levels and all sizes. We should seek the minium aggregation that allows prediction of changes. Not seek to maintain stability. It’s a fools errand.
To mainstream economists that is thinking sideways. But it’s more likely that it’s the correct line of thinking.
Monetary policy is a dull knife. Government of our empire needs to be much more granular and tactical.
Rob
Jul 22 2010 at 1:31pm
After visiting with some professors who teach PhD macro, I got the impression that the current situation is a failure of microeconomic regulation or bad government policy. The DGSE paradigm is alive and well.
fundamentalist
Jul 22 2010 at 1:35pm
Hayek criticized what became DGSE in 1937 in his speech “Economics and Knowledge”. He wrote:
available at http://mises.org/resources/88
As Hayek wrote, the assumption in equilibrium theory before DSGE models was that the economy acted as a single individual would act. DSGE modelers ignored Hayek and made the assumption explicit in their misguided frenzy to tack on some micro econ feel to their macro. In equilibrium, a single individual never has trouble coordinating consumption, savings, investment and production. That’s why equilibrium is possible. But in the real world, you have to explain how many separate individuals can have the same perfect foresight, because perfect foresight is necessary for perfect coordination and equilibrium to exist.
Depressions are caused by a lack of good foresight on the part of many business people at the same time. Yet DSGE has assumed away the problem, which is why the models can’t predict depressions.
Doc Merlin
Jul 22 2010 at 5:27pm
@fundamentalist:
Part of the problem is that microeconomic choices are inherently transitive, but macroeconomic and public choices are frequently (and often) intransitive. Arrow’s Theorem covers the public choice intransitivities, but I haven’t seen a good treatment of macro intransitivity.
fundamentalist
Jul 22 2010 at 5:36pm
Doc, that’s interesting. The main problem Hayek focuses on is knowledge. The definition of equilibrium is that everyone has subjective knowledge of the plans of everyone else necessary for them to carry out their own plans, plus they have objective knowledge that matches the real world. In other words, knowledge must be accurate and all plans must coordinate. If some have inaccurate knowledge, their plans will fail, and if they don’t know the plans of others with which their plans must coordinate, those plans will fail. Of course, neither assumption in equilibrium is remotely true of the real world.
Mike
Jul 22 2010 at 7:59pm
Your link to Simon’s link is inadvertently set to the Solow article. Can you reset it.
David Wiczer
Jul 22 2010 at 11:34pm
Solow’s fascination with representative agent DSGE models is remarkably narrow. This is 30 years past the state of the art, theory and computation have both progressed far beyond the simple models that he damns. It seems pretty silly to say a science fails because of its primitive state long ago.
In particular, macro models with explicit conflicts of interest, bargaining, income and wealth distributions, default and many other important aspects have been analyzed computed and matched to real world economies. We can introduce distortions both through reduced form ‘wedges’ as John DSGE notes, and microfounded structures. I’m pretty disappointed that Congress heard this rather poor representation of our science and even more disappointed that Solow decided to push the message.
Arnold Kling
Jul 23 2010 at 6:43am
Mike,
Link re-set. Sorry about that
Thanks for pointing it out.
fundamentalist
Jul 23 2010 at 9:27am
So the models did a good job of predicting the current crisis and providing recommendations for fixing it?
fundamentalist
Jul 23 2010 at 9:28am
The acid test of models is out of sample prediction, also called validation. How well have DSGE models passed this test?
Comments are closed.