Hi egleesc,
I can think of two reasons:
1) (as erikasami also indicated) a loopback address is a logical interface. Meaning: it cannot go down (it is always up).
2) Security: by choosing (i.e. part of Design) a special block of addresses to assign loopback addresses, you then know traffic is coming (or originating) for a special purpose - so you can adapt these in firewall rules, networking management triggers etc... hence, also limit the routing possibilities for this block (i.e. you can only reach certain destinations and/or be reached via certain IP subnets)
Combining to the reasons above: "You should be scared if your network management system one day informs you that a local loopback address just went down... "