I used to love the idea of this sci-fi-cum-reality tech stuff, but the fundamental (computer) problem of identification is probably the largest in this no-hope system. Being able to reliably identify ICBMs as opposed to fighter planes, space rockets etc would be virtually impossible - make it too sensitive and it gets loads of false positives (hooray for airliner passengers), make it too weak and a single false negative could spell colossal disaster. A useful balance between the two will likely be impossible to achieve with current techniques and even if it were attainable, it'd be very simple to fool because such a system would rely on very specific assumptions about what equates to a threat.
And I'm in total agreement about its usefulness in today's world, it may have been relevant in Reagan's era but is a hopeless waste of taxpayer money now that we're on good terms with China, Russia, et al. I suspect Blair is co-operating with the US on this to keep transatlantic relations cosy, as if they need strengthening any more than they are already. But as for Bush's motivations, I do wonder... it's a frivolous piece of spending considering the other host of problems the US is facing ATM