The Economics of Roadside Bombs

[mhanson]

You put quite a bit of information together, and perhaps had you used some of your think tank’s unclassified info and stats, you’d end up with a real first class product. What did your colleagues think of the paper ?

JIEDDO has asked me to do a follow-on study to confirm the results. They told me that nobody had yet studied the effect IED countermeasures have on non-IED attacks, and were quite interested (for fairly obvious reasons) in the possibility that this was an area in which nobody (including them) had previously recognized their contributions.

They are also inviting me to present my paper at the Military Operations Research Society meeting at the Coast Guard Academy in June -- anyone here going to be there?

I’m not so sure we ‘missed’ or ‘didn’t look at’ the benefits and effects of jamming equipment.

There certainly have been studies of the effectiveness of jammers before, but I don't think anyone has yet done a quantitative assessment of their impact on things they weren't intended to affect -- non-IED attacks.

While I also agree with your conclusion (the use of jamming equipment surprised and stymied the insurgents activities), I’m not sure how you came up with 1,504 less attacks.

The 1,504 IED attacks figure is a statistical estimate from the correlation between trigger use and IED effectiveness. First, I control for changes in trigger use the insurgents decide to make -- that is, since the insurgents know at least approximately how effective they were last month, how does that affect their balance of triggers used this month? Once that decision is modeled, other changes in trigger use are, I argue, a function of jammers (since more jammers mean less radio control trigger use) and some element of random chance. Those changes in trigger use are correlated with lower IED effectiveness, and from that correlation you can then derive an estimate of how many IED attacks were made ineffective.

Obviously, it's ridiculous to defend the 1,504 figure as opposed to, say, 1,503 or 1,505. It is, in fact, highly likely that the true figure is not exactly 1,504. However, 1,504 is my best estimate, and the probability of being right goes down drastically the farther away you get from 1,504.

If someone's life, or hundreds of millions of dollars, rested on the question of whether the number of attacks made ineffective was 1,504 or 1,503, my paper would be of virtually no help in making such a decision. However, suppose the decision was about how much money to spend on jammers versus equipment that could reduce the effectiveness of non-IED attacks (say, more counterbattery radars). In that case, I think my paper would have something very interesting to say -- that we now know, with high probability, that you get a lot more bang for your buck than we have previously thought if you spend money on jammers. I don't think my paper has any direct applicability to anyone operating below the brigade level, but I think it has some interesting applications for assessing the relative merits of various systems that might be deployed.

[Stan]

JIEDDO has asked me to do a follow-on study to confirm the results. They told me that nobody had yet studied the effect IED countermeasures have on non-IED attacks, and were quite interested (for fairly obvious reasons) in the possibility that this was an area in which nobody (including them) had previously recognized their contributions.

Sounds great ! Good luck with your presentation at the USCG Academy !

If someone's life, or hundreds of millions of dollars, rested on the question of whether the number of attacks made ineffective was 1,504 or 1,503, my paper would be of virtually no help in making such a decision. However, suppose the decision was about how much money to spend on jammers versus equipment that could reduce the effectiveness of non-IED attacks (say, more counterbattery radars). In that case, I think my paper would have something very interesting to say -- that we now know, with high probability, that you get a lot more bang for your buck than we have previously thought if you spend money on jammers. I don't think my paper has any direct applicability to anyone operating below the brigade level, but I think it has some interesting applications for assessing the relative merits of various systems that might be deployed.

I'm no economics expert, but have managed DOD budgets in excess of seven million and well below the brigade level. In short, your paper needed larger numbers to support an initially huge procurement. If we break those figures down to a single soldier's life versus the purchase of a single jamming device, the numbers quickly support the purchase:
Payment to soldier's beneficiary = $200,000 - $250,000
one each MILSPEC Signal Jamming Device = $60,000 - $85,000)

Now all we have to do is get a congressman's son into EOD and the rest of the procurement(s) will go like Sierra through a goose