Texting bad politicians out of office: Experimental evidence from Indian elections

Citation:

Siddharth George, Sarika Gupta, Manoj Kumar, and Yusuf Neggers. Working Paper. “Texting bad politicians out of office: Experimental evidence from Indian elections”.
neudc_paper_180730.pdf2.3 MB

Abstract:

Adverse selection to political office is common in many low-income democracies. In India,
24% of MPs have criminal charges, even though criminal politicians have been shown to hurt economic development, worsen poverty, and increase crime. Using a randomized experiment in the context of state assembly elections in the largest state in India, we test whether information constraints and coordination failure among voters are relevant to this bad equilibrium. We partnered with three major Indian telecom companies to conduct a mobile-based voter information experiment across more than 3,800 villages. Mobile subscribers in treatment villages received voice and text messages informing them of the number and type of criminal charges against all major party candidates in their constituency, or election-related messages without criminality content. We find that voters respond to the content of the information provided – in areas where candidates are revealed to have severe charges, votes for those candidates drop by 7.6 percent and votes for candidates revealed to have no criminal charges increase by 6.7 percent and on average. As a result, overall turnout increases by 1.6 percent in these areas.
Last updated on 08/30/2018