I present a paper at NYU’s development economics workshop tomorrow. it’s an odd paper for me, but by far the most interesting one to write so far.

The basic question: why is there so much variation in the age that armed groups recruit? Take a look at this figure, which shows the age profiles in 12 armed groups where we have survey data:

You can click on the figures to enlarge. The boxes represent the 25th to the 75th percentile of recruitment age. Some groups, like Uganda’s LRA, are almost wholly children. Others, like Sierra Leone’s CDF, are almost all adults.

What’s more, the more children that rebels recruit, the more likely it is they use coercion. Here is what (I think) may be the first cross-rebel regression, relating child recruitment and abduction:

Why? My evidence fro Uganda suggests that, when coerced, children are the most likely to be indoctrinated and disoriented, and so are the least likely to escape.

Here is a graph of self-reported allegiance in the LRA by age of abduction:

And here is the average length of abduction, but age of recruitment:

How do we think about this as economists? I’m working with Bernd Beber (a hot job market candidate in poli sci) to incorporate coercion and indoctrination into standard agency theory. The industrial organization of rebellion.

We’re trying our best to keep in mind the advice from this post.

If our theory is right, we might be going about combating child soldiering all wrong. A draft paper is here.