Thoughts on Panics, from Kim Zussman
Like courage, is panic unidentifiable without hindsight? Perfectly efficient markets ought not to over-react, but there is plenty of evidence that they do (not the least of which is you reading this). The problem is telling at what point sell-offs are overdone (and will be reversed over some period), and when are not (and won't reverse in a meaningful period, e.g. Nasdaq 03/2000 top).
Obvious panics should not occur, because anyone who "knows" that a sell-off is a panic would buy and the opportunities would go extinct.
The debate between counters and political traders is about who can discount, better than the market, the impact of current and future events. You can look for panic in time-series patterns and claim to know as a result of numerical precedent. Or you can try to define panic in hyperbole (media, magazine covers, gas rationing, Muslims, etc) that rarely gets it right.
One way not to settle this is by looking at who was right this week, or last week, or last year, etc. because everyone gets it right sometimes. Furthermore, were any one category of analysis superior by dint of p&l, all the brainpower would flock to it and the advantage would arb away.
Maybe it's best just not to panic.