Do exiled children become tomorrow's invasion force?

Here is a sketch of a model describing the military inclinations of refugee children as they mature. It is relevant to Rwanda, which has been invaded once by a generation of adults who grew up in exile and is now possibly threatened by the same demographic, and Burundi whose conflict since the 1990s may have involved the grown-up children of former refugees. The idea was based on the social stratification results in the paper mentioned in the last post.

The model assumes that the refugee population size is stable and people replace themselves. That is, every woman has exactly two children by a single paired man. A proportion of the first generation is assumed to be willing to invade their homeland, with the rest of the first generation being unwilling.

We can model transmission of the attitude to children in various ways. If having either parent willing to invade means that the children are willing to invade, then being willing to invade is, in maths terms, an absorbing state. The number of people in the following generation willing to invade is no less than in the previous generation, and there is a positive probability that it will rise, so we expect invasion in the next generation to be more likely. Moreover, the probability is bounded below, so that over multiple generations, it becomes almost certain that everyone would want to invade. If having a non-invading parent means that the children would certainly not want to invade, we obtain the opposite results, where an expected decline in militarism would occur.

We can allow for other features to better reflect the observed situation:
- children averaging or randomly selecting their parents' attitudes,
- a drift to indifference where militarism falls over time,
- deskilling of children away from home (for example losing agricultural knowledge of their parents) leading to increased militarism,
- hostility from the country of refuge,
- incitement by leaders,
- UN intervention,
- a utility maximising procedure to microfound the decisions.

Judging from the social structuring results mentioned previously, it seems plausible that for different parameter values, we could see either complete militarisation, complete demilitarisation, or a combination of the two.