Dr. Dani Rodrick points to an article by Oeindrila Dube and Juan Vargas. (Dani Rodrik's weblog: The economics of civil strife). The paper concludes that when prices of labour intensive commodity rise - strife reduces a benefits are wide spread. Whereas when prices of capital intensive commodity rises - strife increases if benefits are concentrated. This has some lessons for poverty alleviation programs.
I am tempted to go to my example of snakes-and ladder model for poverty. Here snakes are poverty creators and ladders are poverty alleviaters. Ladders are typically income related and snakes are expense related.
In the context of ladders, poverty alleviation has two-fold objectives. First is to have ladder-accessibility to all (income growth / economic growth of labour intensive activities). Second is to have more ladders (income generating activities supporting /surrounding the capital intensive sectors)
A quick view on the strife part of paper
In my view, strife actually happens when both the above conditions are negated. Or if income generating opportunities are not available to large scale population while being selectively available to few. AND whatever other income generating opportunities are available are not growing. When oil prices increase in Columbia did not translate into ancillary economic activity then strife is definite.