Wednesday, January 21, 2009

Causality and cure

This Econbrowser post states: "How you think we might get out of our current economic problems has something to do with how you think we got into them in the first place."

This implies that the cure is determined by its cause (or causes). So if there are multiple causes we would treat all of them. But by some accounts e.g. Acemoglu/Johnson there are various contributory factors (not necessarily causes) e.g. the Great Moderation made policy makers complacent that also should be considered. When there are many factors and causes do we address them all equally or are some more important than others? How do we determine the size effect of these factors/causes? Should contributory factors get a smaller weight than "causes"?

And what if there are feedback effects, for instance, if mark to market were a feedback effect, should we "short-circuit" this feedback loop? Or what if credit downgrades caused feedback effects, for instance as in AIG?

Finally, how do we model all these when models have been discredited? Do we return to structural equation models?

Update: Eswar Prasad and Brad Setser explore the global roots of the current financial crisis. If global imbalances are a proximate cause that created the liquidity for the subprime bubble then is capital controls the answer?

No comments: