Matt doesn't think that cash handouts can lead to long-term growth. A cynic might argue that neither does conventional aid. I also agree that targeted handouts could encourage dependency and moral hazard – that's why you'd need either conditional transfers like PROGRESA, or else just completely random/universal transfers.
In the nick of time, one of my very favourite irregular macroeconomist bloggers, Younotsneaky! links to a slightly old paper on the use remittances by individuals in Bangladesh. Much of it goes on investment, which of course potentially means an impact on growth. Of course remittances might not be the same as a completely free cash handout, there might be some moral obligation to use them more wisely, but the figures are interesting nonetheless.