There comes a point in every crypto-currency's life where a major hack threatens the health of the system. It is always possible for major stakeholders (miners) to hard-fork in order to correct the problem, but hard-forks are messy and ugly.
The options for a network are:
a) never reverse transactions, to hell with the share price
b) permit bailouts with consensus
If the BTSX on BTER had been stolen, what would we do? It would certainly be within our power to reverse it with a single update pushed to the delegates. It could potentially "fork" the network if the delegates disagreed with the process.
Given that forks are "difficulty but possible" it sets a certain threshold that must be reached before it would be considered. I think that it may be best to recognize that sometimes a network needs to come to consensus about this stuff and design it in ahead of time so that there are no hard forks.
Fortunately we have delegates and thus we can design a process something like this:
Step 1) Pay a large fee to propose reallocating funds from a set of addresses to a new set of addresses.
Step 2) Delegates have 48 hours to approve the reallocation during which time the funds are frozen.
Step 3) Require 51% of the delegates to approve it.
Step 4) Like the pay-rate, delegates can campaign on a platform of always voting NO and this campaign promise can be enforced.
Step 5) The amount in question must be greater than X% of the shares, and the fee should be very high.
1) Someone could attempt to "bribe the delegates" by proposing a massive reallocation to the delegates as part of approving the process....
a) someone could also do this to bribe miners, forgers, etc to mine on the fork
Bottom line is this: if it is possible to implement via a hard fork and there are cases where people would choose to hard fork, then perhaps we should formalize the process and prevent the hardfork and overall disruption. The mere presence of such recourse is likely to prevent many large thefts in the first place.