Did Jovenel Moise's assassination begin in 2010?

In 2010, Jovenel Moïse was not a political figure. He was a businessman in the North with likely zero political aspirations. 11 years later, he made history as Haiti's first presidential assassination in over 100 years. Despite his anonymity in 2010, could the seeds of the assassination had been planted then?

I'm going to summarize my research, recently published in Contemporary Economic Policy (ungated version), which points to political problems created in the wake of the 2010 earthquake. These problems possibly had a butterfly effect that eventually resulted in Moïse's death. But the research itself does not make this grand claim. It argues a much smaller point: that the foreign intervention in Haiti's 2010 presidential election reduced turnout in the 2015 election. After summarizing the work, I'll fill in the pieces that are too speculative for academic research.

The Earthquake Election

It's no surprise that foreign powers intervened in Haiti's 2010 election. It took place months after a 7.0 earthquake struck just outside of Port-au-Prince, destroying the capital and causing around 200,000 deaths. In such disarray, Haiti needed support. But what is surprising is that the intervention was so brazen.

Since the Duvalier dynasty fell in 1986, foreign powers have always been in the background of Haitian elections. The primary role has been to monitor election-day behavior. Sometimes they have had to intervene, but usually it was soft intervention. The most significant election intervention before 2010 was in 2006. The final vote tally had a significant share of blank ballots, which the constitution said had to be counted. The problem was that the electoral council did not know if the blanks were accidental or intentional protest votes. If they were protest votes, then counting them as such would bring the leading candidate's vote share just below the 50% threshold required to win without a second-round vote. Not wanting to worry about another round, the foreign powers saw a loophole in the constitution: it mandated the blank ballots be counted, but it never said how. So they recommended distributing the blank votes to all candidates in proportion to their vote shares. Thus, René Préval won a second term in office. Since Préval was the leading candidate anyway, this was not a terribly controversial intervention.

But 2010 decided to top it. Despite being mired in earthquake recovery, Haiti still held its elections in November 2010. Turnout was low, with only 23% of the country participating. And the first count of the votes had an even race: first place was Mirlande Manigat with 31.4%, second was Jude Célestin with 22.5%, and third was Michel Martelly with 21.8%. As we saw with Préval, if no candidate gets at least 50% of the vote, then the top two candidates go to a run-off. Naturally, this would be Manigat and Célestin. But the second round was delayed because of a debate of whether Célestin truly deserved that second place finish. Since he was Préval's candidate, some believed that he had used fraud and state power to boost his vote share. Furthermore, a brazenly biased "audit" by the Organization of American States claimed that discarding some irregular ballot sheets would cause Célestin to drop to third. The resistance delayed the election, and the country was at a standstill.

Enter the U.S. Tired of gridlock, Secretary of State Hillary Clinton flew to Haiti and met with Haitian leaders. The same day, the election authorities announced that Célestin had been dropped and Martelly would advance.

This intervention perfectly fits the three conditions that Shulman and Bloom (2012) claim are most likely to create pushback from voters. First, it was salient, everyone knew Clinton came that day and knew she was representing the United States. Second, it was partisan; it disproportionately affected one party. Finally, it was a state actor; this wasn't a random American, it was the U.S. Secretary of State.

I'll get to the intervention's consequences, but I want to point out the irony here. Haiti followed Clinton's advice, and the final round was between Manigat and Martelly. Manigat was a former First Lady, and Martelly was a crude celebrity with zero political experience. Martelly won the election, which he never could have done without Clinton's intervention. In 2016, Clinton found herself in a similar election which she also lost, where she also blamed a foreign intervention. It was this irony that prompted me to do this research in the first place.

Haiti's 2015 Election

My research interest was whether the intervention affected the next election. Imagine it's 2010 and you're living in a country with a weak democratic tradition. In spite of all the troubles of voting, you decide to vote for your preferred candidate, Célestin. Then, after a visit from one of the most important Americans in the world, your government announces that your candidate has been kicked out of the race. Do you really want to go through the hassle of voting in 2015?

There's plenty of evidence that the answer was "no." Turnout in 2015 was 18%, and that's only counting registered voters. Kolbe and Muggah (2016) surveyed over 2,000 households, and the second most popular response for non-participation was a belief that there was no point in voting (the most popular was concerns over fraud). The survey also revealed that households were not excited about foreign interventions resolving election problems.

All of this leads to the natural question: did the 2010 intervention reduce voter participation in 2015?

I argue that it did. If you want all of the details, you can go to the paper. But the basic empirical strategy assumes that the intervention affected Célestin's 2010 supporters more since their candidate was removed from the election. These voters would be the most discouraged from participating in 2015. Consistent with this hypothesis, the analysis shows that sections that voted more for Célestin had a greater fall in turnout in 2015.

But what about fraud? If Célestin's 2010 votes were unusually high because of ballot stuffing, then we would naturally see turnout fall in 2015 because that fraud had been eliminated. While this is mechanically true, there are several reasons to believe it is not the main explanation. First, there was never definitive evidence that Célestin cheated, or (more precisely) that he cheated more than any other candidate. In fact, the second round election between Martelly and Manigat had just as many complaints about fraud. Second, I tested the sensitivity of the results to assumptions about fraud. The conclusion is that for fraud to have a significant effect on the results, then the fraud must have been much larger than anything we actually observed.

I don't believe this is irrefutable evidence for a discouragement effect. But I think it's strong enough that it should be taken seriously. So how does this connect to Moïse's assassination?

The Butterfly Effect

One clear effect from the intervention was that Martelly advanced to the second round and became president. Without the intervention (presumably) the second round is between Manigat and Célestin. While it's hard to imagine counterfactuals under a different president, it's clear that Martelly's poor leadership contributed to many of the problems Haiti faced. Here's just one example: Haiti was shut down in 2018 by protests over embezzlement in Martelly's administration. I already wrote about how that had a tragic effect on entrepreneurial activity in Haiti.

But another effect could be Moïse being in office in the first place. Moïse won the presidency in 2016 in an election he was not expected to win. A key role in his victory was low voter participation. Did the intervention discourage enough Célestin supporters to enable Moïse's victory? Would the frustrations with Haiti's government had been mitigated if a non-Martelly administration was in office?

Again, counterfactuals are hard to imagine. But Martelly's administration generated discontent. And his administration was not held accountable, allowing his chosen successor to take office. Could the tragedies had been avoided if the U.S. had not intervened? Or would we be writing about a completely different set of problems caused by not intervening?

Foreign policy is hard.