In February, South Sudan’s president, Salva Kiir, swore in the country’s new vice president — rebel leader Riek Machar, his longtime rival. The ceremony marked a renewed hope for peace in South Sudan. As they had done in the past, Machar and Kiir smiled for the cameras and referred to each other as partners, urging forgiveness and pledging commitment to the still unfinished peace settlement. This time, they insisted, they would work together to end the country’s ethnic civil war and bring stability to the region.

Such ceremonies can signal unified commitment to a peaceful future. But seeing a leader whose supporters have targeted one’s ethnic group for violence then switch gears and endorse a peace deal might raise suspicions — does the leader really support the deal, or will there be more violence in the future?

Our research indicates that people living in conflict settings tend to be suspicious about endorsements of peace deals. In our survey, we found that when it was revealed that Kiir or Machar had endorsed a tentative peace proposal, people who belonged to groups victimized by forces aligned with the respective leader were less supportive of the deal.

How we researched trust in peace deals

Research from places that are not experiencing conflict shows that people tend to trust leaders from their own ethnic groups more than they trust leaders from another ethnic group. We wondered if trust works the same way in a conflict setting like South Sudan. Because the sides fighting in South Sudan’s civil war are divided largely along ethnic lines, we suspected that people would be particularly trusting of messages from leaders who shared their ethnic identity — we call these “in-group” leaders. And we thought they would be distrustful of leaders from other ethnic groups — the “out-group” leaders — that had targeted their communities for violence.

So it’s logical to expect that people would be persuaded by endorsements of peace deals coming from leaders who share their ethnicity — but be more distrustful of peace deal endorsements from leaders outside their own ethnic group.

To explore these questions, we took advantage of a brief lull in violence in 2016 to conduct the first-ever endorsement study of peace policies in an active conflict setting. Following South Sudan’s independence in 2011, Kiir, who is ethnically Dinka, became the new country’s first president. Machar, who is ethnically Nuer, became vice president.

In late 2013, Machar fled the capital, Juba, to become the de facto leader of the opposition, marking the beginning of the civil war. Over the course of our nearly six-month study, he returned to the capital to resume the role of vice president. But he fled again in July 2016 and fighting broke out, and we ended data collection.

We conducted nearly 1,000 surveys with respondents from four different ethnic groups: the Dinka, Nuer, Luo (who in our sample faced indiscriminate violence from forces loyal both to Kiir and Machar) and Shilluk (who faced targeted violence from forces loyal to Kiir). Our respondents came from 42 villages across eight counties and five of South Sudan’s 10 states.

We first asked respondents some basic questions about their background: their gender, ethnicity, exposure to violence and more. Then we asked respondents to use a scale of 1 to 5 to rate their level of support for six different peace policies — with 5 indicating the highest support. We chose policies from actual tentative peace agreements between Kiir and Machar, mostly from the 2015 peace accords.

We investigated the effects of leader endorsements on peace deal support by randomly assigning respondents to either see no endorsement before each policy, or a Kiir or a Machar endorsement.

We then compared peace deal support among respondents who received no policy endorsement with support among those who received endorsements from either Kiir or Machar before ranking their level of support.

Respondents didn’t trust endorsements from an ‘out-group’ leader

Consistent with our expectations, we found that endorsements from leaders of another ethnic group backfired. Support for all six peace policies, and the peace deal overall, dropped precipitously when a respondent read that an out-group leader from a group with a history of violent attacks on their specific ethnic group had endorsed that policy.

For our study, this meant that a Kiir endorsement led to drops in support among the groups that Kiir-allied forces had targeted: Nuer and Shilluk respondents. Similarly, a Machar endorsement left Dinka respondents less in favor of the peace policies. We also found some evidence that decreases in support for the peace policies were more severe among those who had faced more violence, and thus had the greatest reason to distrust the leaders they viewed as responsible for that violence.

And we found that leader endorsements had little effect among respondents whose communities were not targeted by one side for violence. We found that peace deal support among Luo respondents was unaffected by endorsements from either politician, for instance — this was the group that faced indiscriminate violence.

But endorsements from ‘in-group’ leaders weren’t persuasive

Contrary to expectation, we didn’t find that a Kiir endorsement shifted support among the Dinka respondents (who share Kiir’s ethnic identity); nor did a Machar endorsement shift support among Nuer respondents (Machar’s ethnic group).

Why? Our interpretation is that in a context of prolonged conflict, where leaders have repeatedly promised and failed to deliver peace, people become distrustful of all leaders. While an in-group leader’s endorsement might not make someone suspicious of a deal, people might not view the endorsement as a particularly valuable source of information.

Leader endorsements can actually result in lower levels of public support for peace agreements, our findings suggest. That doesn’t mean leaders should not endorse peace deals. But our findings indicate that increasing intergroup trust and building safeguards into peace deals to protect against a resurgence of violence — and communicating those safeguards to citizens — may help increase public support for peace deals, and the prospects of lasting peace.

Nicholas Haas is a PhD candidate in the Politics Department at New York University. His research looks at causes, consequences and potential solutions to intergroup inequality, polarization and conflict.

Prabin Khadka is a PhD candidate in the Politics Department at New York University. His research covers insurgency, counterinsurgency and peace-building with a focus on the Horn and sub-Saharan Africa, the Middle East and Nepal.