"Press 1 for Roads": Mixed Evidence on Improving Political Communication

Co-authored with Saad Gulzar and Luke Sonnet

To investigate whether lack of information is one reason politicians may be unre- sponsive to voter preferences, we conduct a randomized control trial with senior politicians in Pakistan that collects preferences using Interactive Voice Response (IVR). IVR allows politicians to script questions for voters and voters to respond using cell phones. There is strong interest by politicians in soliciting opinion via IVR; additionally, relatively high numbers of voters respond. Nonetheless, politicians fail to use the information provided to modify their on-the-ground engagement with voters or types of service delivery. We also do not observe improvements in voter evaluations of politicians or in electoral support for them. A forecasting exercise shows that experts find these outcomes unexpected. One implication is that information may not be the primary determinant of the gap between what voters want and what politicians do?

paper_public241224.pdf

Download PDF • 1.32MB