Pat Gunning raises the question of the definition of a shock in 
macroeconomics. My understanding is that a 'shock' is some unexpected event 
(unexpected, that is, by market participants), or a change in expectations 
(which, by definition, cannot be expected). Expected events are already 
built in to plans, so do not shift any curves away from expected (planned) 
positions. Unexpected events make people change their plans, which may be 
represented by shifting curves. Pat Gunning's examples (and Kevin Hoover's) 
all seem to fit this pattern.

One more specific comment: Pat Gunning wrote: 'Bursting an asset bubble: 
... The money that disappears from the asset market and that causes the 
asset price to fall must be put somewhere else. It must cause other asset 
prices to rise'. This seems to me to be wrong. Asset prices can fall (or 
rise) with hardly any actual transactions (just enough to register the 
price). No money has gone somewhere else - the asset values have simply 
been written down. In a (rational) bubble, the high (subjective) 
probability of continued rising prices balances a low chance of a fall. 
When the bubble bursts, the low probability of a fall becomes the knowledge 
that the fall has actually happened, shifting expectations radically.

Tony Brewer