Women walk past campaign posters of Ugandan President Yoweri Museveni in Kampala in February 2016. Museveni won reelection for a fifth term as president last year. (Carl de Souza/AFP via Getty Images)

Elections are a hallmark of democracy. So why do dictatorships around the world also hold them? While autocratic elections are often characterized as “sham elections,” with the official vote winner clear beforehand, elections in fact have systematic and substantial effects on the durability of dictatorships.

In a recent article in World Politics, we examine 389 elections in 259 dictatorships. We found that elections can actually prolong dictatorships in the longer term. But we also found that holding elections is very risky for dictators in the shorter term.

Why dictators lose sleep before election day

In 1986, the official count of a snap election in the Philippines stated that President Ferdinand Marcos won, by a fairly narrow majority. Filipinos and international observers cried foul and proclaimed his opponent Corazon Aquino the true winner. After mass protests, Marcos stepped down and the country welcomed the regime of “Cory” Aquino.

In February 2016, Ugandan President Yoweri Museveni may have feared a similar fate. To preempt large-scale coordinated protests on election day, his regime blocked Twitter and Facebook, arrested the main opposition candidate and deployed security forces. Museveni officially won the contested vote and remained in power, but the regime’s actions suggest a non-trivial probability and concern that events around election day might spiral out of control.

Elections are a double-edged sword

Elections often bring regime-threatening protests or coup d’etats — and that’s why they’re so dangerous to autocratic regimes. But autocrats also gain something from holding elections. There is a long-standing debate among political scientists on whether elections are, on balance, stabilizing or destabilizing for dictatorships. We argue that they are both, but with a key nuance: Elections confer long-term benefits. The regime can co-opt members of the opposition, for instance, or learn more about the strength of the opposition. Elections also help dictators build a strong organizational apparatus and signal their strength to intimidate potential opponents.

The flip side, as the Marcos and Museveni examples suggest, is that elections also can produce short-term instability by enabling opposition groups to coordinate their actions right around when the election takes place. Elections are thus a double-edged sword for autocratic regimes.

We analyzed 259 autocratic regimes

Drawing on data from 259 autocratic regimes across the world, from 1946 to 2008, we test these concepts, by analyzing the dynamic relationship between autocratic elections and the risk of regime breakdown, using different statistical models.

We find clear evidence that the period around and right after an autocratic election — and during the election year in particular — is associated with a greatly increased risk of regime breakdown. In fact, 50 percent of the dictatorship deaths we looked at occurred in election years, although election years make up only 22 percent of all dictatorship years.

This short-term destabilizing effect remains when we try to account for the possibility that elections are held, or postponed, for strategic reasons by clever autocrats, such as Museveni, who anticipate that elections can affect the regime’s survival. Our statistical models suggest that the risk of a regime breaking down in an election year is about five times higher than if the last election happened two to five years earlier.

Why would dictators want to hold elections?

There’s also some evidence, although not as clear-cut, that holding elections makes for more stable dictatorships in the long term. We think this is an important underlying reason that autocrats like Marcos take the gamble of holding elections, despite the risks to the regime around Election Day.

Here’s how these long-term benefits played out in Mexico’s autocratic regimes from 1929 to 2000. The party in power, the Institutional Revolutionary Party, or PRI, for decades managed to conduct elections without seeing major threats to its grip on power.

In fact, studies by political scientists argue that the PRI used these elections as a device to prolong its rule. The party used elections to selectively co-opt supporters — but deter opponents by displaying organizational strength and broad public appeal. These stabilizing effects continued long past election day.

Are there lessons for opposition parties?

Our research indicates that opposition actors in autocratic regimes may find a unique window for dissent around election time, when autocratic incumbents are particularly vulnerable. After World War II, dictatorships across the world have held elections, on average, every fifth year. Yet, every second dictatorship that broke down did so during an election year. Since elections are held at specific and commonly known dates, the analysis suggests that they offer opportunities for regime opponents to coordinate their actions and remove an incumbent autocrat.

Carl Henrik Knutsen is professor of political science at the University of Oslo.

Håvard Mokleiv Nygård is a senior researcher at the Peace Research Institute Oslo (PRIO).

Tore Wig is a postdoctoral fellow in the Department of Political Science at the University of Oslo.