Can democratic regimes in unequal societies redistribute in favor of the poor without partisan strings attached? Under what conditions are politicians willing and able to tie their own hands in favor of rules-based implementation of social programs? This dissertation manuscript presents the analysis of the policy trajectories of conditional cash transfers in two prototypical clientelistic polities: Mexico and Brazil (1997-2006). I develop an original theoretical model of choice for program implementation in two dimensions (rules vs. discretion and centralization vs. decentralization), and posit that the choice for the implementation model of direct transfers for the poor –as a choice between clientelism and rules-based implementation- results from the electoral calculus of the federal government, which is a function of the relative brokerage capacity (partisan territorial presence) of the governing party in the regions where the poor are overrepresented. I combine process-tracing methods with the statistical analysis of program allocations at the municipality level and find no evidence of political influence in the selection of beneficiaries in both cases. My findings suggest that the Mexican federal government invested in credible non-partisan targeting of the Progresa-Oportunidades and a highly centralized model of implementation as a way of tying the hands of local PRI brokers, who could otherwise be disproportionately favored through the extraction of clientelistic rents from the political manipulation of the program. In the case of Brazil, I find a similar trajectory towards program depoliticisation that warrants, however, a different explanation: the relative brokerage capacity of the federal government in deprived regions also provided for incentives for depoliticisation (with a decentralized implementation model), but not directly as a function PTs territorial presence, but the territorial implications of the governing coalition of Lulaå«s first administration. By 2006, Bolsa Fam lia and Oportunidades stood as emblematic examples of the ability of the state to channel resources for the poor without partisan strings attached, and remarkable poverty alleviation programs. This research projects provides for a politically-informed understanding of these choices.