It’s no surprise that India—the world’s largest democracy—is a hub for policy research. But Ford School assistant professor Yusuf Neggers studies the subcontinent for both personal and practical reasons.
“My mother is from Uttar Pradesh,” Neggers explained. Hired by U-M just last year, he was a highly sought after academic in his field, according to Ford School Dean, Dr. Michael Barr.
Uttar Pradesh was also the site of Neggers’s first field study for an ongoing candidate criminality messaging project, which he launched during his time as a policy PhD student at Harvard’s Kennedy School. Part of the funding for this salient project now comes from U-M.
“Two classmates from grad school and I learned how high the rates of these serious criminal charges are in Indian politics. We were shocked, and it motivated us because we thought this intervention had potential to have some real bite to it.” He was further inspired by Milan Vaishnav’s 2017 book, When Crime Pays: Money and Muscle in Indian Politics.
Neggers and his co-investigators—researchers Sarika Gupta and Siddharth George—traveled to India on multiple occasions to meet with telecommunications companies and an implementing NGO partner. They collaborated with these organizations to disseminate texts and automated voice messages with candidate criminal charge information in the immediate run-up to election day.
“We wanted to make sure that we, as researchers, weren’t just extracting data without providing some sort of value add,” he said.
This sort of mindfulness is why—when a telecommunications partner expressed concerns of political backlash the night before the first of the project’s four planned phases, which included the actual circulation of messages—the researchers agreed to cancel the remainder of the experiment once the initial round of voice and text messages went out, even though the proposal had included three additional sets of electoral races.
“From their perspective, it makes sense to err on the side of caution. We’re really small clients, and if they get punished it’s a big deal,” Neggers said.
But, following the release of criminal charge information to voters the next day, there was no retribution from elected officials. Neggers and his collaborators hope to conduct similar trials in the future and feel that this demonstrated absence of pushback will be encouraging to potential partner companies.
Candidate criminality statistics are available online, but many voters may have difficulty accessing this information, especially those in rural areas. Sending these messages directly to voters improves the likelihood that they become knowledgeable about this important dimension of candidate profiles.
“As it turns out, if you reveal severe criminal charges to the population, on average you do see vote shares for those candidates drop. For people revealed not to have any charges, you see an increase in their vote shares. These results are consistent with a setting where, for some part of the population, these information constraints are relevant.”
As for future implications of his research, Neggers is hopeful for a change in the composition of who ends up running for office.
“If parties know these types of candidates have lower chances of being elected, there is the possibility that they will select against people with particularly heinous criminal charges,” he said.
“There’s been a hypothesis that voters, on a larger scale, already know whether candidates have criminal charges and just don’t care. Now, we see that what might be going on for some part of the population is, if you’re a rural voter, you might not have a good source of information about candidates’ criminal profiles. We think these results are quite encouraging.”