Matt makes an important observation:

“In judgments about how much effort to put into securing nuclear stockpiles around the world, I repeatedly see gross underestimation, not overestimation, of likely terrorist capabilities.”

This is a fair description of how people in many countries see the threat, and it is a real impediment to progress on securing weapons and materials. But we typically have the opposite problem in the United States.

When politicians tell Americans that the sky is falling – and they often do just that when it comes to nuclear threats – American believe them. If you want evidence, just look at how many people, before the Monday NIE, thought that Iran was about to get the bomb.

Matt is also right to be distressed about the fact that many foreign leaders completely dismiss the possibility that terrorists could make a nuclear bomb. If they read any one of dozens of books and studies written over the last thirty-five years (including mine as well as several that Matt has written) they’ll learn that it is indeed quite possible. Again, it’s American people and policymakers who I find tilt too far in the other direction – while possible, building a bomb is, in many cases and under many circumstances, far from straightforward, especially if a group wants to keep its conspiracy under wraps.

Matt challenges me to provide a real case where a proposed defense that was worth the money was rejected because it was imperfect. Let me suggest two elements of one case that show this dynamic at work. (As I wrote in my original comment, I was providing a bit of a caricature – but I do believe it gives a feel for how many people think.) That case is precisely the one that Matt identifies:

“Congress’ rush to require that 100% of the cargo containers coming to the United States be scanned for radiation by 2012 – with absolutely no statement as to how good that scanning should be, no similar requirement for better detection for any routes other than cargo containers, and no system-level assessment of how useful this would be – is an obvious case in point.”

I agree with much of this criticism. The demand for 100% screening, accompanied by no statement as to how good that scanning should be, is a great example of where an obsession with perfection and a focus on the worst case distorts strategy.

Imagine that I proposed inspecting only twenty percent of incoming containers, and supported that by arguing that in the context of a broader system, anything more would likely do one of two things – it would steer terrorists to a different approach, or convince them not to risk attempting a nuclear plot – and that as a result, buying any more scanners would be a waste of money. We all know what the response would be: people would ridicule the scheme for letting four-in-five terrorist plots succeed. Believe me, this is not a theoretical exercise – I’ve tried.

The obsession with worst cases also sheds light on why we’re buying expensive detectors that still fall short of perfect. Every audit mandated by Congress seems to focus on the fact that whatever detector it’s studying can’t detect pure weapons-grade uranium metal. Well guess what: unless a terrorist group makes a mistake (something that is entirely possible but has nothing to do with fancy radiation detectors) that’s pretty much a losing cause.

On the other hand, there are a host of other plausible nuclear materials (as well as intact weapons) that are easier to detect. Instead of pouring huge sums of money into schemes that might slightly increase the odds of success against one worst-case target – what Congress tends to demand – we might focus on simpler, cheaper technologies that would be effective only against a large but limited set of smuggled materials and terrorist schemes. It’s sensible logic, but try convincing people that they shouldn’t aim for perfection, and you’ll have trouble getting very far.

On a broader note, Matt worries that the balance of focus in my book will tilt the reader in ways that the detailed analysis might not support. In particular, he is concerned that in trying to fill in what I see as massive gaps in what others have written, I’ve given the misimpression that security at the source should not be an urgent focus for policy. I’m sympathetic to his concern – it’s something I thought about while I wrote the book.

I suspect that part of the imbalance he sees is the result of genuine disagreements on how effective different parts of the defense can be, and part does reflect my desire to focus on offering a serious take on some big holes in our thinking about nuclear terrorism. (An earlier draft of the book included an explanation of why I was spending relatively little time on describing measure to improve security at the source – the reason what that others have done a great job already.) I encourage everyone who reads my book to also read the authoritative takes on what to do about security for nuclear materials that I recommend in the text.