Two Traditional and New Views on Banking Panics

It is better to be cautious about characterizing one historcal event with statistical tests. We set null and alternative hypotheses in limited views.

There have been two traditional views on banking panics One is the sunspot-based view (Diamond and Dybvig 1983). In this view, random withdrawals could be triggered by depositors with self-fullfilling motivation. And the other is the information-based view (Gordon1988). In this view, bank runs could be triggered by some signals on expected return of bank portfolio.

Many empirical studies have tested these views, and the resuls show that fundamentals could explain bank suspension or financial crisies well. Then, is it ridiculous to to worry about random-withdrawals?  Can we say that the random-withdrawals  view is wrong?

Some papers develope Diamond and Dybvig (1983)'s model using global game methods while most papers support the information-based views before the Crisis of 1998 or of 2008 ( Goldstein and pauzner 2005; Morris and Shin 2006). In this new view, random withdrawals could be triggered by fundamentals shock. This means that some banking panics, which have been regarded as information-based events, might be characterized as some self-fullfiling activities. Contagion effects during the panic or interbank activites come to be explained by fundamental factors.

Theoretical researches could evolve as well as empirical studies. There is no need to stick to two opposing theories. We are able to, (or have to) provide more evidence for new theories. This is so simple. On the other hand, when some paper show the results of hypothesis tests, we often forget this simple thing.