No, I don't agree with that. Systems can perfectly exist in an unstable equilibrium, until something pushes them away from it - that would be an external shock. Even the crises that are generated from within the system needed a push to actually become crises. I agree that the roots of recent crises (and most crises, come to that) lay within the (imperfect) systems themselves, but I don't agree with the view that they can evolve on their own, without any triggers. So, no shocks - no crises.

I agree with the first statement: "systems can perfectly exist in an unstable equilibrium, until something pushes them away from it".

What I am saying is that shock not a necessary condition. Not at all. They can be generated without any triggers. And you, as a physicist know that

So, why would you go around and blame some out-of-the-model(world) forces for every crises? I think we should allow for the possibility of internal crises and try to device policies to avoid them. As long as we treat them as external, the best thing we could do is to device policies to make economies more resistant to those extra-model forces, which can strike from any side in any guise.

I still think that even in physics we need a trigger. Take a ball placed safely on the very top of an inverted-U shaped hill. It is an unstable equilibrium - any force applied to the ball will cause its movement away from that equilibrium point. But you need a trigger, blow of wind, earthquake, something, to apply this initial force. Or take overcooling - you can cool distilled water to the temperatures way below zero, but once you introduce the freezing center - the whole system instantly becomes frozen. I can't recall any examples of such phenomena, not needing an external push. If we were to talk of the quantum world, it's even better Everything there is probabilistic - and, with a certain degree of simplification, there's your white noise processes, affecting those systems.

My point is that even internal crises need an unanticipated push. Because, if the push is anticipated, it is already part of the model and hence cannot create a crisis (very simplistic, I know, but still).

I have not said we do not need a trigger. I said we do not need a shock (as understood in modern economics) to serve as the trigger.

And I agree with the ball example, just as I agreed before that external shocks can indeed generate crises. No doubt about that. What I am saying, again, is that it is not only external shocks that can do this.

There are many models in physics which display drastic changes in aggregate behavior (the crisis) at some point in time without any external intervention. These are, for example systems that display one type of behavior as long as certain values (which are endogenously changing) stay within certain bands, but display sudden phase changes at point when they leave the bounds (clearly the trigger is crossing the band, but its not a shock). For economics applications of few of these you can have a look at early works of Axel Leijonhufvud of UCLA. In general models with self-oganizion criticality from physics fit the bill. There triggers are not external and there are no internal shocks either, in a sense that there is nothing out of ordinary going on with the variable. But, system suddenly collapses.

Now imagine how much more exciting (and useful) would be the economics discipline if we could have models like this. How much easier it would be to design the policy: it would have a clear aim of keeping a certain variable within the defined bands. Instead at the moment economic policy making looks like asking questions: what type shock should we expect? when is it likely for this chock to occur? how can we resist the shock of certain type? how marge of a shock can we actually survive? And, sadly, we usually never have answers to firs two questions out of these four.

In short, you want to present an economic system in a way similar to thermodynamic one? I can't think of another example from physics at the moment that would fit your description (and I'm not sure of thermodynamics either ), please tell me if you have anything specific in mind. But if we think of thermodynamics, self-organization there is an aberration with an extremely low probability, right? And the triggers there actually, if I still remember physics' stuff correctly, cause the complete disruption of the system - not a crisis magnitude that would be useful for us. I've heard of this Axel guy (from you, most likely), but haven't actually looked into his research, could you send/link me whatever you think would be interesting to me?

As for the models, answering all the four questions - I think that human factor is too high in economics and it can only be captured by shocks.

Thermodynamics is one option. There are many more. Think of percolation models, where slight changes in density have crucial consequences in aggregate outcomes. The same is true for network models. Think of work that Thomas Schelling got the nobel prize for. If crucial parameters in these models could become endogenous variables (and I believe they can) we would have something valuable at hand.

I have written a bit more extensively on these issues here.

Leijonhufvud is just one of the guys advocating this approach. Most of his arguments should come through by reading this. It might well have been the "EconBlogs" that you have heard from about Leijonhufvud, rather than me: there was a discussion past october/november that Nobel should have gone to him, rather than to Sargent! Anyway, there are more guys in the camp: another notable one is Peter Howitt of Brown. The most notable european one is Giovanni Dosi.

As about the human factor: it is exactly its "highness" that pushes us to distance ourselves from non-human/extra-model forces in forms of shocks and embrace self-generated crises.

I do not know why it did not work, but in the previous post it should be:

...by reading this.

Do we have a macro model that incorporates financial pyramids (that eventually exhaust themselves), stock market casinos, excessive leveraging, and other kinds of crisis generating behaviors?

Basic macro models actually prevent Ponzi schemes from being created, by imposing a transversality condition, which basically states that you can't borrow more than the present value of your lifetime earnings stream. You can introduce excessive leveraging, as far as i know.

There are sunspot (self-fullfilling prophecies) models in macro, so yes such models do exist, and they have arrived in the mainstream.

Although stock market casinos? That makes me *shudder*.

Don't shudder, Michael. Stock markets are worse than casinos (or better, depending on which side of the "table" you sit).

I personally participated (only once, in Israel, many years ago, so won't be prosecuted) in a well-orchestrated effort to drive up the price of an option by 1000%!!! I was told by an uncle (who was close to the epicenter of the conspiracy) to buy at 50 cents and sell at 500 cents, so there was no risk for me. (Unfortunately, I was a poor student at the time and did not have much money to spare/invest). People who were late to join this little pyramid have lost big time. This type of criminal behavior is actually rational (exploiting other people's greed and naivete), but has nothing to do with market valuation of a firm.

In my case, the crime was perpetrated by a few people in a tiny market. Big institutional investors or the Goldman Sachs of the world can and do the same on a much bigger scale, generating crisis after crisis, at fairly regular intervals. These crises and the intervals at which they occur could be modeled. For instance, they must be governed by the people's ability to forget (which depends on the magnitude and impact of the latest crisis) and availability of other investment or saving opportunities (for instance, interest on bank deposits, government bonds, rate of return on investment in own business or education).

How useful are the currently available models, with or without sunspots, in modeling these types of rational behaviors, and the crises they generate?

Indeed, sunspot models are discussed even in grad macro texts, such as Blanchard and Fisher for example.

They are, although to be honest I am very sceptical about the field of econophysics, and its contributions so far.

As are majority of mainstream people "We" are still probing grounds. I believer there is a potential.

The idea comes from as early as Georgescu-Roegen, but still hasn't fulfilled it's potential. I'm not sure, whether it's possible at all