> Trying to apply that sort of economic analysis to a statistically nonexistent phenomenon like major terrorist attacks is utterly useless by comparison. It doesn't matter how accurate or inaccurate our assessment of the costs is, because we have no idea how likely they are to occur in the first place, and no framework to assess what the cost of preventing them would be.

That's complete bunk! What's so special about terrorist attacks that make them so hard to model, as compared to other things governments routinely model like inflation, technology transfer from NASA programs, etc? An absolutely ludicrous assertion.