Since I’ve had three emails in one week asking me about this issue, I figured I might as well blog about it and have something to refer people to instead. The questions have all been variants of:
· Are women better remitters than men?
· Does having mothers migrate result in worse outcomes for kids than having their fathers migrate?
· Do women receiving remittances spend them on kids, while men waste them on booze?
There is certainly many studies which claim to answer questions like this. For example, Guzman et al.  look in Ghana at how the sex of the remitter and the sex of the household head receiving remittances is related to household expenditure patterns; Abrego  compares Salvadorean families with immigrant mothers to those with immigrant fathers; Lindley  remarks that, in Somali communities, it is sometimes said that ‘women are “better” remitters than men (even that it is better to have one daughter abroad than ten sons); Battistella and Conaco  look in the Philippines at impacts on children left behind of having mothers versus fathers migrating, etc.
So what’s the problem? It depends on what the question really is asking. If it is a simple descriptive question (e.g. do the women who choose to migrate send a higher or lower amount of remittances home than the men who choose to migrate?) – then there is no problem – this is a factual question that can be addressed by simply looking at the data (indeed in a study of African migrants in the OECD  my co-authors and I found male migrants remit more than female migrants).
The problem is when one wants to move from descriptive to causal statements. If our questions are instead “does having a man migrate rather than a woman migrate cause less remittances to be received? Or cause kids remaining in the home country to do better at school?” then we face very difficult identification problems.
The problem in identifying the impact of gender on migration outcomes is that ideally to answer this question we would like a lottery which randomly chooses which one of the mother or father migrates, and which one stays behind (and if we really want to separate the effects of the gender of the recipient from those of the gender of the parent remaining with the children, also randomize whether people are in same-sex or opposite sex marriages!). Otherwise we worry that the types of households that send women abroad are quite different from those that send men abroad – for example, often one hears discussion that women migrating abroad leads to more divorce or to kids doing worse in schools – but it might be that women are only choosing to migrate when they don’t get on so well with their husbands, or when their kids are doing so badly in school that they decide they need the extra money to pay for private school.
Identification issues plague most of the migration literature, since it is hard enough to tell a reason why a particular household engages in migration and an otherwise similar household does not – aside from migration lotteries, people have looked at a range of instruments like economic conditions abroad or historic networks to try and get at this issue. Even if one was convinced by these instruments, there are two problems using them for gender analysis: first, the instruments often don’t provide variation within a household – e.g. if the village has a good migrant network in an overseas destination which is receiving a positive shock, this should apply to both the man and the woman in the household; and second, even if one could get gender-specific instruments, we would at best identify LATE impacts - and so it is not clear that it is that meaningful to compare e.g. the impacts of migration for women who will migrate only if Chicago has had a positive employment shock and they live in a village with a historic network of women in Chicago to the impact for men who live in the same village who have a historic network in Los Angeles and who will only migrate if Los Angeles has a positive economic shock and otherwise won’t migrate.
Now this isn’t to say that I don’t think it is plausible that gender plays an important role in the way that migration impacts on households, just that I don’t know of any study to date that has convincingly been able to identify what this impact is. Anyone know of a study I am missing?