Working Paper: NBER ID: w26925
Authors: Abhijit Banerjee; Nils T. Enevoldsen; Rohini Pande; Michael Walton
Abstract: In 2010, we informed a random set of Delhi councilors, some ineligible for re-election in their current ward, that a newspaper would report on their performance shortly prior to the 2012 city elections. Using slum dwellers' spending preferences, we created a councilor-specific index of pro-poor spending. Treated councilors increased pro-poor spending in high-slum wards. Cross-cutting experiments suggest that the public nature of report cards, not access to information on public services per se, incentivized councilors. Data on party ticket allocation and electoral outcomes shows that, in low-information situations, credible public disclosures of politician achievements matters to both parties and voters.
Keywords: Public Information; Politician Accountability; Elections; Delhi; Experimental Evidence
JEL Codes: H40; O12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| increase in propoor spending by councilors in high-slum wards (H53) | alignment of spending with preferences of slum dwellers (H53) |
| influence party ticket allocation (D72) | treated incumbents being more likely to run for reelection (D72) |
| anticipated public disclosures through newspaper report cards (G24) | change in voter preferences (K16) |
| anticipated public disclosures through newspaper report cards (G24) | increase in propoor spending by councilors in high-slum wards (H53) |
| anticipated public disclosures through newspaper report cards (G24) | increase in assembly and committee attendance (D72) |
| anticipated public disclosures through newspaper report cards (G24) | influence party ticket allocation (D72) |