We investigate the relative merits of unconditional cash transfers (UCT), conditional cash transfers (CCT), and the effects of improvements in education quality on efficiency and welfare. In our setting, some parents underinvest in their children's education because capital market imperfections prevent them from borrowing. Under sufficiently accurate targeting, CCT are more effective than UCT in enhancing the efficiency of these households' decisions. However, UCT is superior to CCT in terms of welfare unless targeting is perfect, in which case UCT and CCT are equivalent. Education quality is welfare improving, but may not be efficiency enhancing when public education quality is very low.