Asset issuers can default at any time without needing the exchange to go down. In GLBSE's case the lack of any provision of investors' details provided a convenient window for issuers to vanish/claim ignorance (or do a Gigavps and charge a hefty fee for recognition of your claim - pricing smaller non-US investors out of their holdings). That excuse has gone with BTC.CO and (to an extent) Bitfunder (the asset list there is incomplete - as it only includes investors who opted in by providing a BTC address or ticking a box to allow sharing of their email address) but the capability for issuers to vanish is ALWAYS present and, without enforcement of contracts, can't be seriously mitigated.
All the time there are issuers there will be issuers who default. The rate of default/failure has to be factored in when deciding what rate of return is acceptable for investments (over-simplifying, if 10% of assets default/fail per 3 months then you have to make 11% profit per 3 months on the 90% just to break even - so anything paying less than ~ 4% per month should be ignored unless you have good reason to believe it's in the 90% not the 10%. Those numbers are illustrative not intended to be taken as fact - actual default/failure rate may be lower or higher: my money's on higher, dependent on how you define default/failure.)
This of course runs into the problem of the interest vicious circle (as expected returns go up to compensate for rate of failure actual businesses are selected against, actual scams are selected for).
Because it has about two assets with any value in them.
How is that different to MPEX ?
Different because MPEx does not also have a bunch of worthless cruft cluttering it.
Different because MPEx has other stuff besides stocks (like options).
Cryptostocks also has options.
"To have" does not mean "there is the word written on a web page".