Politicians and Poverty Reduction

A long line of work in political science centers around the topic of clientelism, also known as vote buying. The idea is that politicians (the patrons) provide goods, services or benefits to the public or to those not in positions of power in exchange for votes. (Such favors, for example, could involve the allocation of a large development project to a swing-locality or district). There are lots of reasons to be displeased with "vote-buying" but a big one concerns the marginal returns to public expenditure: funds that are allocated for political reasons may not be allocated productively from the standpoint of some other societal goal. For example, if we are interested in poverty reduction, public funds should be allocated to those who are actually poor (yes, I am oversimplifying here!). Obviously, there is no guarantee that funds allocated by a political formula will follow this more technocratic logic.

So are we destined to an equilibrium where the incentives of politicians and the public are not well-aligned? Not so, according to two recent studies which examine the political impacts of conditional cash transfer programs (CCTs), which are schemes that provide cash to poor families who meet certain objectives (i.e., attending monthly check-ups, sending their children to school, etc; see here for a beautifully detailed account of CCTs the world over) and have been shown to have had numerous positive benefits on health, schooling and general circumstances faced by the poor. The first, by Yale political scientist Ana de la O, looks at the impact of the Mexican CCT, Progresa. Progresa was initially rolled out to a randomly selected subset of localities before more universal rollout a year and a half later. De la O uses this variation to show that people randomly exposed to Progresa longer are around 5 percentage points more likely to vote for the incumbent.

In the second study, Marco Manacorda, Ted Miguel and Andrea Vigorito use a regression discontinuity approach to identify the political impacts of Uruguayan CCT PANES. Comparing individuals just on either side of a pre-designated eligibility score (based on a composite of household and individual socioeconomic characteristics), they find that cash transfers lead to a whopping 21-28 percentage point increase in the probability of voting for the incumbent!

What I like about these papers, besides the use of program evaluation methods to identify causal impacts on voting behavior, is that it suggests that there is a better equilibrium out there: politicians can get re-elected on the basis of policies that have demonstrated large and positive effects on human development rather than through doling out public funds in potentially unproductive ways. My hope is that this catches on in other parts of the world...