I would change it to remove the panic consensus part as that is quite subjective and open to gaming.
That would destroy the entire security model! Then you are trusting the entire reserves to 101 individuals only. The gaming is hopefully solved by the tax on their stake if they call a false panic.
I'm imagining a future where thousands of heterogenous organisations could be operating within the BitShares system, and many private and common-use block-chains could be employed by by that organisational network, with the need to transfer and communicate between them **. A general problem to deal with will be that of delegated authority. Delegated authority is a common organisational problem where authority (and with it trust) over certain assets (or even votes and other intangibles represented on a block-chain) is delegated to another person or entity to manage. Traditional examples could be shareholders delegating to managers, managers delegating to subordinates, members delegating to boards, or the public delegating to their voted representatives. It would be useful to have flexible tools that could allow enough variation to deal with a wide number of preferred decision structures.
In that context, it seems the idea of a panic consensus is one method for dealing with the more general problem of managing delegated authority. But is the panic consensus described the most generalised form, or is it a specific form of a more general group of settings of ensuring security in cases of delegated authority? If the latter, what are the more basic ingredients? I have no view on whether it is or is not, but it would be beneficial if the settings of the IUAs allowed as much flexibility as possible for organisations to implement whatever security structures they can conceive of and agree to.
** (as an aside, it would make an interesting conversation at some point to think about how block-chain technology will most likely be implemented by society in the future, as that can lead us to design the appropriate tools and structures)