We take issue with the reasoning of Coniglio et al. (2009) that whereas better-skilled illegal migrants will prefer to return-migrate, lower-skilled illegal migrants will not. We argue that under asymmetric information, all the illegal migrants are initially paid a wage based on the average productivity of the group of illegal migrants. The better-skilled illegal migrants thus face two "taxes:" being paid less than if they were legal, and being averaged down. Therefore, better-skilled illegal migrants can be expected to expend more effort to become legal than lower-skilled illegal migrants. And once legalized, there is no reason for the better-skilled illegal migrants to want to return to their country of origin more than the lower-skilled illegal migrants. Thus, it is the lower-skilled illegal migrants that are likely to dominate the return migration flow. We argue that in other respects too, the model of Coniglio et al. is not based on reasonable assumptions, and that even under the postulated assumptions, the model suffers from several inconsistencies. In particular, when the rate of return to savings is an increasing function of skill level, we would expect there to be few better-skilled individuals among illegal migrants in the first place. Also, an obvious distinction between savers and borrowers is ignored.