This couldn't backfire, could it?... Europe-Libya refugee agreement edition

There's some classic unintended consequences coming soon from the new agreement between Italy and Libya on preventing refugee migration to Europe. Jalel Harchaoui and Matthew Herbert report in the Washington Post:

The seas off western Libya have been quiet since late July. Before that, they swarmed with smugglers’ boats overfilled with migrants, mostly sub-Saharan Africans heading for Europe. From 23,000 migrants per month, the flow of arrivals has slowed to a trickle.

The migrants are accumulating on Libya’s coast and many are incarcerated in opaque circumstances. Their movement has been stymied by militias, who have turned on the northbound flow of migrants they once profited from. Deep in the southern desert, emergent militia groups evince the goal of closing the border with Niger and Chad to migrants moving north — attempting to patrol areas that none of Libya’s three rival governments ever secured.

Motivating the Libyan militias’ newfound zeal for blocking migrant movement is a new policy spearheaded by the Italian government and embraced by the European Union. The approach relies on payment to militias willing to act as migrant deterrent forces. Italian government representatives use intermediaries such as mayors and other local leaders to negotiate terms of the agreements with the armed groups. They also build local support in the targeted areas by distributing humanitarian aid.

What happens when you offer to pay Libyan militias for detaining migrant refugees before they reach Europe? Harchaoui and Herbert suggest that it empowers nonstate armed groups, stunts efforts at building credible security for the future, and may breed conflict. All of that is probably true. But they fail to recognise one of the other unintended consequences, which relates to the incentives the payments create.

When you offer to pay Libyan militias for detaining migrant refugees, you get more detention of 'migrant refugees'. That doesn't sound bad - it sounds like what you intended. Until you take a closer look at who the 'migrant refugees' are who are being detained. Maybe some of them are genuine refugees who were intending on migrating to Europe. But catching those refugees may be costly for the militia (especially if the militia are being paid by the refugees to help them get to Europe). Better then to find some non-paying disadvantaged group and round them up to be detained. Or perhaps, strike a deal with some group who can pose as refugees and split the payment with the militia? There are many ways for the militia to game this system, where the refugee flows towards Italy would barely reduce, while the payments to the militias greatly increase. This is Sudanese slave redemption all over again (see my post on that topic here). Or Project Phoenix (but probably without the killing).

One alternative may be not to pay the militias for the number of migrants they detain, but for the reduction in the number of migrants actually reaching Italy. Although, if you do that you might not want to find out how the militia reduce the number of refugees reaching Italy. And you'd have to find some way of allocating the payment between competing militias. And there would be a free riding problem, since militias might then get paid even though they were doing nothing to reduce refugee flows. Aligning incentives without unintended consequences is hard!