Book Talk: Why Do Bloodstained Groups Win Postwar Elections?
In the latest Talking Policy episode, Sarah Z. Daly, an associate professor of political science at Columbia University, talks about her book Violent Victors: Why Bloodstained Parties Win Postwar Elections. This interview was recorded on October 27, 2022. The transcript has been edited for length and clarity.
Your new book is a comprehensive study of elections after civil wars and answers the question: why do so many voters support parties that engaged in mass violence against civilians during wartime? Talk to me about what inspired you to want to study this very perplexing phenomenon.
Several personal experiences inspired me to write the book. As an undergraduate at Stanford, I took a human rights course in which we learned about Chile’s coup and the repressive dictatorship that took over 40,000 lives.
I assumed, given the human rights abuses of Pinochet’s regime, that all Chileans naturally rejected Pinochet. And then I went to live and study in Chile, and I lived with a family that was very much pro-Pinochet. The family joined the ranks of 44 percent of Chileans who were pro-Pinochet, even after the transition to democracy. I was struck by this.
I was struck again while I was researching my first book. My book asked, why do some armed groups silence their guns while others re-militarize and return to organized violence? And for this book, I engaged in 18 months of fieldwork in Colombia, conducting surveys of ex-combatants, their families, psychologists, and the civilian communities in which they lived, engaging in participant observation across the armed groups, and interviewing over 300 former combatants, commanders, victims, and civil society leaders. Victims told me about the massacres, the rapes, the torture, and the kidnappings that the paramilitaries had carried out.
And yet I found that in many places the populations tolerated and even supported the former paramilitary forces and their allied politicians, even after the paramilitaries had demilitarized.
This political behavior exhibits itself broadly. For example, in El Salvador in 1994, the party tied to the death squads, responsible for the vast majority of political killings, won the elections. In Colombia in 2018, the party of Alvaro Uribe, facing hundreds of investigations for ties to paramilitaries, in what Human Rights Watch called one of the worst episodes of mass atrocity in the Western Hemisphere, won the elections.
Around the world, populations vote for political parties with deep roots in the violent organizations of the past. They do so in the aftermath of nearly every civil war, both long and short, in ethnic and non-ethnic societies, in rich and poor countries, in the presence of peacekeepers and in their absence, in nearly all regions of the world. This puzzle is what motivated me to write the book.
Your book combines case studies from Latin America with experimental survey evidence and new data on postwar elections around the world. You have reviewed 205 belligerent parties that transitioned out of war between 1970 and 2015. So you’ve looked at a huge amount of experience across countries, across wars, across elections to try to understand this really strange phenomenon of people voting in groups that committed violence against them. What did you find?
Overall, I found that the electoral success of bloodstained parties depends not on the extent of their atrocities or electoral coercion, but on the military outcomes of the war. Bloodstained parties, if they win the war, successfully campaign as the best providers of future societal peace. Winning belligerents are able to claim credit for peace, which serves to justify their use of atrocities, and translates into a reputation for competence in the provision of security. Then, if through its campaigning, the belligerent party can signal to voters that it will restrain its violence in the future, it can own the security issue that is paramount to large numbers of voters emerging from the anarchy of civil war.
War outcomes are powerful predictors of the electoral performance of belligerent parties, both rebel and government. If they won the war, belligerent parties perform well, even where elections are free and fair.
Observational survey data from 16 countries showed that an average of 54 percent of citizens were most concerned with securing the future. And these security voters were significantly more likely to cast their ballots for the winning combatant party over either the militarily losing or non-belligerent parties.
What is a really emblematic example of this phenomenon?
Guatemala is a good example. Guatemala’s 36-year war was brutal. The military under FRG [Guatemalan Republican Front] party leader Ríos Montt carried out the Mayan holocaust, which qualified as genocide under international law. The Truth Commission found that government forces, under Ríos Montt, committed over 90 percent of the war’s atrocities. The war ended with a government win. In 1999, free and fair postwar elections were held that pitted Ríos Montt’s party against a strong party without blood on its hands, PAN [National Advancement Party]. PAN was formed by technocrats and business people and had no involvement in the military government. The losing, far less violent rebel group, URNG [Guatemalan National Revolutionary Unity], also tried its hand at the polls. The untainted and the less tainted parties lost those elections.
Guatemala’s electorate was victimized—one-fifth were direct victims. They cared about security. Thirty-four percent cited security as the most important issue, a proportion rivaling all economic issues combined. And 89 percent of the electorate were unaligned, co-optable voters. So, to perform well, the parties had to appeal beyond their core supporters to capture the votes of these victimized security and swing voters. PAN pinned its security competence to its lack of past transgressions and argued that the FRG would yield only “more impunity and abuses of human rights.” To make this legalistic claim, PAN ran clean civilians in suits, targeted moderate voters, and emphasized punishment of the FRG’s violent past.
In response, FRG adopted the optimal strategy for winning belligerent successor parties: it ran as a “Restrained Leviathan.” FRG argued that Ríos Montt had brought peace to the country. It claimed that its credibility on security lay with its ties to the coercive apparatus and the symbolic edge embodied in its strongman Ríos Montt, recognizable as “the old-fashioned caudillo, the man on horseback who saves the nation.” Ríos Montt campaigned only in Army uniform as the party’s symbol of strength, and argued that Guatemala was emerging from anarchy which justified submission to a strongman.
FRG also had to signal restraint. So Ríos Montt’s candidacy was paired with leftist Portillo, a victim of the conflict, who could signal that FRG would not revictimize the former armed and unarmed political left. FRG undertook further costly signals of restraint by purging human rights abusers from its elite and stacking its legislative ticket with mostly civilian candidates, including its former enemies.
The UNRG rebels, meanwhile, lacked the ability to offset their violent past with credit for peace, and so had little credibility as providers of security in the post-civil war period. The URNG rebels, therefore adopted an optimal electoral strategy for losing belligerents—tactical immoderation. They apologized for their transgressions, appointed only civilian candidates, and advanced more ideologically distinct positions to compensate for their security competence disadvantage.
In line with the book’s theory, voters counterintuitively credited the brutal FRG for peace, blamed the less brutal UNRG rebels for the wartime violence, and saw the non-belligerent PAN neutrally.
FRG was able to win 85 percent of security voters. Fifty-nine percent of the population swung to FRG and it won 47 percent of victims. Only 36 percent of the victims voted for the non-belligerent party.
Do voters have an accurate idea of the violence that a party meted out during the war? Is it that they aren’t fully aware of the violence that was committed? Or do they vote for violent parties because they’re afraid of them?
So, there’s definitely a fog of war. But I argue that, for the postwar period, that fog is clearing. Many of the elections took place after truth commissions had publicly reported [on violence]. I have an information experiment in my survey in Colombia and talked with victims at length. Voters knew who the perpetrators were. As one of my government interviewees expressed in El Salvador, “It’s not as if people didn’t know what had happened. Everyone knew.” Across my case studies, the evidence suggests that voters had an accurate idea of the facts of the violence committed. But war outcomes produced asymmetry in the attribution of blame for that violence.
Among the hundreds of cases you looked at, were there exceptions that are worth noting?
There are three types of aberrations. First, there are several cases that defy the trend of high-security salience ahead of the founding elections. These are cases in which, for example, the wars took place in noncontiguous or satellite territories—so the violence only affected certain regions and minority populations. This was the case, for example, in the Philippines, the United Kingdom, and Indonesia, where the general population didn’t acutely feel the civil war violence taking place in the island territories of Mindanao, in Northern Ireland, or in East Timor [respectively]. So security voting was not a powerful force in their postwar national elections and the victor parties didn’t get a significant boost from war outcomes.
Then there are cases in which the rebel and government belligerent parties ran poor campaigns and were punished electorally for doing so. So, for example, the KLA [Kosovo Liberation Army] ended up losing to the non-belligerent Democratic League of Kosovo Party because the KLA successor party failed to moderate beyond its wartime constituency or signal restraint.
And then there’s a final set of cases in which the brutal war winner loses the election because voters choose non-belligerents who offer the rule of law over promises of iron-fist security. For example, in Liberia, despite military successes, neither of the rebel groups nor the belligerent government party emerged as significant forces in the 2005 election. Instead, two non-belligerents—a female technocrat and a football star—competed for votes because the citizenry was disillusioned with short-lived episodes of past postwar peace and preferred civilian parties.
So there are exceptions. But overall, the theory that your book explores does hold across decades, across countries, across cultures. What does this mean for governance? What kind of governments emerge from all of this? Are they good governments? And if people are betting that the party that won the war is best equipped to secure peace, is peace what they get?
Elected war-winner governments prove able to not only tie their own violent hands, but also to deter the losers from re-militarizing. As long as the balance of military power doesn’t change after war, there exists little reason for either the war winner or the war loser to re-initiate violence because a new war wouldn’t have a different outcome, and so peace tends to hold. This is the most likely outcome after post-war elections. But if the balance of power inverts after war, and the electorate chooses the now weaker war winner, electoral results end up becoming misaligned with military power, and the newly empowered belligerent has incentives to return to war.
How good are newly elected postwar parties at securing public goods?
The book suggests that citizens often sacrifice social welfare for security. For example, in Colombia I found that when a paramilitary mayor won an election, the municipality experienced on average an 85 percent reduction in thefts, but a 17 percentage point reduction in educational coverage and a 61 percent reduction in educational spending compared to when such mayors narrowly lost the election.
The mechanism, and I take this up in greater depth in a British Journal of Political Science article, is that among coercive politicians, prioritization of security crowds out resources for other public goods and social welfare.
What about justice? Having a group that committed heinous crimes against civilians win an election feels really unfair.
It’s true, the election of violent victors forestalls legal accountability for perpetrators of crimes against humanity.
But this mitigating role of security on desires for retribution against war winners may wane over time, and so demands for prosecution may expand. This means that perpetrators are likely unable to evade blame forever.
So there’s promise that over time, progress toward transitional justice may become possible. Are there examples of that happening?
Across the world, we see lots of cases in which transitional justice is stalled in the short term, but becomes possible over the long term. Ríos Montt is a very real example of that—he later faced a genocide trial. And you see this with Fujimori in Peru, long after the end of the war, being tried for some of his human rights abuses. There are definitely trends that justice becomes more possible over time.
Your book illuminates some very painful trade-offs and perverse incentives—for example, that violence in war, if it helps a warring faction win the war, may give them the upper hand electorally. Or that a warring faction might have a perverse incentive to start a war if victory is likely, if that means that electoral victory is also likely. What do you take from all of this? And you worried about warring parties reading your book?
I am concerned, since I call them bloodstained parties, that I will never again be able to do research in places where I’ve done fieldwork. But I want to clarify the issue of perverse incentives. So, the book does seem to imply perverse incentives for belligerents to engage in ruthless violence to obtain the upper hand militarily and then electorally, or dangerous incentives to start wars if victory seems assured. And it indicates that belligerents can escape electoral retribution for their transgressions in the short run. But I want to make the point that transgressions are unlikely to increase prospects for wartime military success. Indeed, there are so many studies out there that have demonstrated the counterproductive nature of indiscriminate violence in war, which I confirm in much of my earlier work. And, in the long run, ghosts from the past tend to catch up with their victimizers. It’s not the atrocities that bolster belligerent parties in postwar elections. It’s winning the war that does.
But you also point to the really painful tradeoffs inherent in transitions. Numerous scholars have noted that what’s necessary to avert instability and recurrent war may perversely protect human rights abusers from justice, prevent the country from effectively resolving what Samuel Huntington called “the torturer problem,” and hinder the deepening of democracy.
Where I fall on this is: it’s voters themselves who ultimately have to weigh these tradeoffs and choose between peace or justice and democratization in the short run.
And if the elections are free and fair, the outcome suggests that, with their ballots, war-ravaged voters tend to prioritize stability. But the good news of the book suggests that justice and democratization may also eventually become possible.
So the question becomes, how can we speed up the normalization of politics after violence? How can we generate conditions for transitional, legal accountability, protect and deepen democracy, establish rule of law, and enhance social development? Interventions aimed at securing the peace, buttressing the balance of power, preventing waves of criminality, reducing the urgency of security issues, and countering efforts to spin the violent past may dampen some of the perverse electoral potency of these war outcomes and amplify opportunities for justice and liberalism.
A survey done on behalf of the Violence Prevention Research Program at UC Davis found that just over 20 percent of Americans who were surveyed were willing to condone political violence, at least some of the time. When asked if having a strong leader for America is more important than having a democracy, more than 40 percent of those polled said that, yes, they agreed in some respect with that statement.
This seems to suggest that, just as you found in your book globally, people are willing to trade certain values for others. We’re in the lead-up to the Jan. 6th committee report release, and I’m curious what you think about how support for political violence by parties or candidates in the U.S. might affect them at the polls.
In recent years, U.S. politics have exhibited chilling similarities to the context I deal with in my book. This motivated me to write a Washington Post Monkey Cage piece on Trump’s use of the strongman electoral playbook in 2020. This playbook has been picked up by others who want to heighten the salience or urgency of anarchy to justify an adherence to a savior strongman. And they argue they shouldn’t be blamed for violence, but instead should be rewarded with votes for preventing greater turmoil. They canvas surrounded by military uniforms and brag about links to violent militias. They use propaganda to spin a narrative in which the strong break the law so as to provide stability.
Given their perceived ownership of the security issue, these politicians have incentives to keep security threats alive through their scary rhetoric, but they may also perversely feel compelled to do so by fomenting threats of actual low-level insecurity and violence. Core voters, and sometimes even swing voters, may respond with their security-motivated votes. This plays into a broader pattern that, when feeling threatened, people prove willing to forego rights and liberties in favor of hardline approaches to security and crime, at times voting seemingly against their self-interest.
This suggests that parties, terrifyingly, may be electorally rewarded for their support of political violence in the U.S. I’m currently engaging in research on what can be done to counter this trend and undermine fear-mongering politics and politicians. Making voters feel safe to reduce the salience of security is important as is bolstering non-belligerent parties and ensuring that they adopt a Rule-Abider political equilibrium strategy—so that means promising security within the confines of the law, running candidates with clean records, advancing moderate policies, targeting the median voter. Outsiders could help to strengthen these parties’ advantages on non-security issues: on economic issues to further boost their electoral successes. Also critically important is countering misinformation and the spinning of reality.
I want to end with a personal question, which is, after many years of research on a topic that interested you long ago in grad school, what has surprised you most, and what are you most encouraged by?
I was surprised to find that those early observations were not an aberration. I went into the project with that same sort of optimism that I had when I first went to live in Chile; I assumed that everyone would reject human rights abusers. But in all of my work, I seek to find glimmers of hope from the sometimes less bad cases. They might not be perfect, but I think they offer a roadmap to greater, more positive outcomes.
At the same time, I think there are positive stories in this—I mean, out of war come elections that do bring peace more than they bring war recurrence, and that do bring democracy, even if it’s not liberal democracy. And there’s hope for greater democratization and liberalism and greater justice over time.
Yeah, and that is also a finding of your book: that wars can end. Sarah, thanks for being with us on Talking Policy.
Thank you so much for having me.
 Of Guatemalans, 88 percent did not have a party affiliation in 1999. FRG was poised to win 86 percent of voters who had cast votes for it in 1995, but also 35 percent of previous PAN voters and 50 percent of those who had not voted in the 1995 elections. Of the population, 59 percent seems to have swung to FRG.
Thumbnail credit: Elena Hermosa; witness testifies during genocide trial of former Guatemalan military dictator, Ríos Montt.