An hour outside of Varanasi, India, the Ganjari village cricket ground is hot and dusty. Birds pick at a cow carcass beside the road, and a stand further down sells samosas. Players arrive on motorbikes, and the men cluster in teams of five around MIT Abdul Latif Jameel Poverty Action Lab (J-PAL) surveyors, who enter details like the captain’s name and the batting order into tablets. It’s a friendly Saturday cricket match—and also a development economics experiment in mediating caste interactions.
“You could think of the cricket pitch as a microcosm of caste issues in India,” says Matt Lowe, the leader of the study. This summer, Lowe traveled to India through the MIT-India program at the MIT International Science and Technology Initiatives (MISTI) to continue his field work on the J-PAL project.
Lowe studied economics as an undergraduate at the University of Cambridge in England and joined the MIT Department of Economics in 2012. During the summer after his third year, he worked with a large Indian carpet manufacturer near Varanasi, a city in the northern Indian state of Uttar Pradesh.
“In retrospect I was quite naive. I had all these ideas, and I would pitch them every day. But a lot of them were really wacky,” Lowe says. “I was thinking of what was interesting to me intellectually, and not what would make sense for the firm.”
One aspect of his summer at the carpet factory that struck him was caste; he observed that, when offered spaces in a paid training program, women from lower castes often refused to participate because they felt that members of high castes in the neighborhood surrounding the training shed would abuse them.
Back at MIT, Lowe began talking with his advisors, Esther Duflo, the Abdul Latif Jameel Professor of Poverty Alleviation and Development Economics and the co-founder of J-PAL, and Daron Acemoglu, the Elizabeth and James Killian Professor of Economics. They envisioned a project in the same region of India that would focus specifically on caste—and how caste interactions might be improved through a series of integrated cricket matches.
“The cricket tournament just gives you this natural story,” Lowe says. “It gives you a very structured way to study and observe social interaction.” Plus, he says, India loves cricket.
Each five-player cricket team orders itself for batting, and chooses bowlers, who are similar to pitchers in baseball. But, every player is not guaranteed to bat or bowl in an over, which is similar to a baseball inning. Lowe would track who was chosen as captain and who was prioritized to bat and bowl. Arguments, collusion, and high fives given would also be recorded. Photos of the team taken before each match would later be coded for who stood next to whom, and body language would be observed.
Together with the Sarathi Development Foundation, a nonprofit organization based in Uttar Pradesh, Lowe and his team approached pradhans, village leaders, and prominent community members to build interest in hosting paid tournaments. These areas typically hold tournaments once or twice a year, but players pay to play, and only the winning team receives a cash prize, explained Mustufa Patel, a J-PAL research consultant who heads up the field operations in Varanasi. To qualify, villages needed to have an even split between the three caste categories: General, Other Backwards Class, and Scheduled Caste/Scheduled Tribe. Lowe asked the pradhans to draw maps of their villages to ensure that the castes lived separately.
Although caste-based discrimination was officially outlawed in India more than 60 years ago, an individual’s caste is often easily discerned by their last name, and can still affect their opportunities, access to education, profession, and social interactions. According to the India Human Development Survey of 2012, 37 percent of upper caste households in Uttar Pradesh practice untouchability, the shunning of lower castes and confining them to menial and despised jobs. Nationally, this figure is slightly lower, with roughly one out of four households confirming the practice.
After a series of baseline surveys ascertained each player’s caste, mapped their social networks among members of the tournament, and assessed their cricket abilities, the players were assigned to teams of same or mixed caste. Players were then randomized to receive payment either based on the performance of their entire team, or their individual performance. With support from the J-PAL Governance Initiative, the Weiss Family Fund, the Shultz Fund, the Center for International Studies, and MIT-India, the project grew to support eight tournaments with more than 140 players each.
“Traditional studies in J-PAL used to have just two surveys. That was it—baseline and endline. In the middle, the partner implements the program,” Lowe says. “But this is much more like you are observing and surveying every single day, capturing every interaction.”
Comprehensively capturing these interactions was made possible through the twenty-odd J-PAL surveyors, the supervisors, and the field manager working on the project around Varanasi. Equipped with a tablet and a motorbike, at least two men attended every cricket match—one to score the details of the game, and the other to observe how players talked with one another, argued with other players, teams, or the umpire, and participated in moments of encouragement, like hugging or bat-tapping. Intern Yashna Shivdasani of Wellesley College’s Undergraduate Research Opportunity Program is classifying anecdotes from the surveyors’ notes into categories such as “comments on monetary incentive” and “caste-based insults.”
For the players, the events were ostensibly just an opportunity to play cricket—and a chance to make as much as 300 rupees, which is not an insignificant sum for players who might expect to earn 200 rupees (about $3) for a day of work. Lowe took care to focus discussion on community building and community interactions within villages when talking about the project with the surveyors.
Eight tournaments down, the surveyors have caught on. Raj Rajesh, who has recorded more than a hundred cricket matches, says when he joined, he thought the focus would be more on cricket than on caste. He has worked as a surveyor for J-PAL for three years, and comes from a caste-segregated village in the northern state of Bihar. Speaking through an interpreter, he explains that when he first started, he was skeptical about inter-caste interaction; now he is beginning to understand that it is possible. And it could be useful, he says.
The last tournament complete, Lowe is now measuring changes in players’ social networks, as well as having them vote on which teammates should attend a professional cricket coaching workshop. Through MIT-India, Lowe is in Varanasi again this summer to complete the final stage of the study, in which players receive a gift of slippers or gloves. But there’s a catch: Players are given two right-footed slippers or right-handed gloves or two left-footed slippers or left-handed gloves and must exchange with one another to make a useful pair. Lowe will be able to track who has traded with whom.
If the effects of the competitions are long term, a cricket-based intervention could be worth scaling up across Uttar Pradesh, Lowe says. Many people have the ability to score matches, and the effort would be relatively low in cost compared to other interventions, even with players continuing to receive monetary incentives. Eventually, entire villages could be randomized, and the effects of mediating caste interactions through cricket could be studied even on people who do not play in the matches, such as women.
“Cricket is something that 50 percent of these guys play every single day. They play it every day, but they still only have one lower caste friend,” Lowe says. “Though they don’t currently play together, as soon as they are asked, they play together willingly. That alone tells you it’s worth trying.”