"Many People Like Me Got a House": Material Benefits and Sociotropic Voting

• In Event: Clientelism and Distributive Politics 2

Abstract

Governments target benefits at swing voters in the hope of winning their votes. But do voters have to personally benefit for preferences to change? Existing work on distributive politics assumes this. We find that people who do not benefit from a rural housing program in India are just as supportive of the governing party compared to beneficiaries. This is because many people in their social network benefit from the program. Our pre-registered study identifies the causal effect of getting a house by leveraging an arbitrary cut-off that separates beneficiaries from those next-in-line to benefit. We interview 530 Dalit households around this cut-off. We find, on a four-point scale, beneficiaries are 0.6 to 1 scale point more likely to think the incumbent has done something for them, 0.4 to 0.87 points less likely to think there is ethnic favoritism or broker discretion in distribution, and 19 to 22 percentage points more likely to think people voted for the incumbent out of gratitude. However, beneficiaries are no more supportive of the governing party than those next-in-line. Strikingly, the incumbent is very popular among those next-in-line, potentially because their social networks are saturated with the benefit. 70% of them personally know at least one other beneficiary, typically nine other beneficiaries. We rule out several alternatives: anticipation effects at the cut point, low satisfaction, misattribution, short-term material shocks, overriding ethnic factors, and clientelistic capture.

Author

• Shikhar Singh, Yale University