Saturday, August 8, 2009
My Lucas Critique
Over at the Mises blog--and note that the Mises blog is different from the Mises Daily article--I have a short post about Robert Lucas' defense of mainstream economics. Lucas basically argued that nobody could have predicted the housing crash, because if they had predicted it by the day before, then it would have happened the day before. And so on. I wrote:
In the comments Troy Camplin argued that, contrary to Lucas, the existence of price bubbles obviously shows the Efficient Markets Hypothesis (EMH) is false. I responded:
This is far too clever and slippery. Someone can evaluate a situation as unsustainable (or poised in an "unstable equilibrium") without being able to predict exactly when the break will occur....
Also...I'm starting to think the efficient markets hypothesis is a state of mind, a consciously chosen way of looking at the world. I'm not sure what it would mean to really falsify it. It seems that any attempt to test it would rely on assumptions about seemingly random events, which in the final analysis would mean you weren't really testing the efficient markets hypothesis alone.
In the comments Troy Camplin argued that, contrary to Lucas, the existence of price bubbles obviously shows the Efficient Markets Hypothesis (EMH) is false. I responded:
Troy,
Your understanding of what the EMH is, is obviously different from Lucas'. Lucas obviously is aware of what happened to housing prices, and he doesn't think it violated EMH.
This is partly what I was getting at in my post. I'm pretty sure that Lucas thinks the housing market was hit with a really big shock in 2006, and that made prices fall.
Lucas knows it must have been an unexpected shock, because...(you fill in the blanks).
So my point is that Lucas thinks he just empirically tested whether EMH held up during the housing boom and bust, and he thinks it passed through with flying colors. Yet what wouldn't pass with flying colors?
Don't misunderstand, I'm agreeing with you: The housing bubble is a great reason that I personally don't endorse the EMH. But technically speaking, the EMH is consistent with what just happened. So Lucas isn't wrong, he just doesn't realize how a priori his worldview is.
Comments:
You need a model of market equilibrium to test the EMH. Your model and Lucas's model must differ. EMH can only be tested this dual-test way. It's a long-made point by Fama. He writes about it on his blog once in a while, recently too I believe.
Sand is continually added to the top of a pile -- does the timing or fact that it will collapse upon itself depend on my predicting it?
No.
Physics tells me I can't predict when it will collapse (we're dealing with dynamic non-linear phenomena here) but is does tell me that the pile will collapse and why.
Many of the most scientistic physics worshipers in economics don't know any physics -- too bad for economics.
No.
Physics tells me I can't predict when it will collapse (we're dealing with dynamic non-linear phenomena here) but is does tell me that the pile will collapse and why.
Many of the most scientistic physics worshipers in economics don't know any physics -- too bad for economics.
It seems that any attempt to test it would rely on assumptions about seemingly random events, which in the final analysis would mean you weren't really testing the efficient markets hypothesis alone.
Test it?
Are we still talking Austrian economics here, where empirical tests can not be conducted?
Lucas basically argued that nobody could have predicted the housing crash, because if they had predicted it by the day before, then it would have happened the day before.
Are we still talking Austrian economics here where extreme aggregation is shunned? If I know the housing market is going to collapse, just how the hell does this translate into everyone knowing?
Or am I, because I have this unique knowledge, supposed to go out and sell short the entire housing market all by myself?
Test it?
Are we still talking Austrian economics here, where empirical tests can not be conducted?
Lucas basically argued that nobody could have predicted the housing crash, because if they had predicted it by the day before, then it would have happened the day before.
Are we still talking Austrian economics here where extreme aggregation is shunned? If I know the housing market is going to collapse, just how the hell does this translate into everyone knowing?
Or am I, because I have this unique knowledge, supposed to go out and sell short the entire housing market all by myself?
RW wrote:
Test it?
Are we still talking Austrian economics here, where empirical tests can not be conducted?
No, we're talking about the efficient markets hypothesis, which its proponents believe is an empirical proposition.
But as I'm arguing in the post, it's actually not empirical. The way Lucas et al. explain away apparent counterexamples, would apply to any possible counterexample.
You're right, I think the main problem with Lucas is that he's taking the Lucas critique too seriously--he's imposing rational expectations on everyone. I.e. the only reason you and I can have different forecasts of home prices is if we have different, private signals. But we have to share the same model of the housing market, and moreover our (shared) model is correct.
So yes, in that framework that Lucas is imposing, not discovering empirically, it's true that any popping bubble must have taken everyone by surprise, in a sense.
Post a Comment
Test it?
Are we still talking Austrian economics here, where empirical tests can not be conducted?
No, we're talking about the efficient markets hypothesis, which its proponents believe is an empirical proposition.
But as I'm arguing in the post, it's actually not empirical. The way Lucas et al. explain away apparent counterexamples, would apply to any possible counterexample.
You're right, I think the main problem with Lucas is that he's taking the Lucas critique too seriously--he's imposing rational expectations on everyone. I.e. the only reason you and I can have different forecasts of home prices is if we have different, private signals. But we have to share the same model of the housing market, and moreover our (shared) model is correct.
So yes, in that framework that Lucas is imposing, not discovering empirically, it's true that any popping bubble must have taken everyone by surprise, in a sense.
<< Home
Subscribe to Posts [Atom]