The resilience of Brazilian democracy, in the face of efforts by supporters of former President Jair Bolsonaro to undermine the validity of the recent presidential election, is a story that has just about everything. Courtroom drama, baseless claims of voter fraud, brawls in the halls of power, and hopeful hints that democracy may not be as fragile as it has seemed in recent years.

There has been, unsurprisingly, a lot of focus on the powerful individuals whose decisions ensured that the election result was respected, such as an aggressive Supreme Court justice and the leaders of its military. Their decisions were undoubtedly important. But focusing on a few people’s choices can obscure another important issue: the strength of Brazil’s democratic institutions — and how that affects ordinary Brazilians’ lives.

I know that “institutions” can sound dry as a topic — a second ago we were talking about riots, now I’m conjuring visions of paperwork and buzzing fluorescent lighting — but stay with me here.

Because I want to talk about a new paper about Brazilian political institutions by Camilo Nieto-Matiz, a political science professor at the University of Texas San Antonio, and Natán Skigin, a Ph.D. student at Notre Dame. It reads a bit like political science as scripted by Martin Scorsese — light on the paperwork, heavy on the murders and gangland politics. And although it is not specifically about Bolsonaro or the recent election, it offers important context about the conditions that brought the country into, and potentially out of, a democratic crisis.

A surprising way to reduce violent crime

Brazil’s electronic voting system has made headlines around the world with the false claims by Bolsonaro, as president, that it was rife with fraud.

But Nieto-Matiz and Skigin began studying the system years earlier, when Brazil first began rolling it out to districts across the country. They noticed that it seemed to be having a surprising effect: When electronic voting was introduced into a particular area, violent crime there quickly fell.

“That was really puzzling,” Nieto-Matiz told me when we spoke last week. They had expected to perhaps find a relationship between electronic voting and particular policies: perhaps a benefit to illiterate citizens, whose votes were more likely to be counted under the new electronic system than the old paper one. But the decrease in violence seemed to happen almost immediately, before any new policies had a chance to take effect. What could account for that?

When they dug a little deeper, they found that the new voting system seemed to make it slightly less likely for political parties that gain votes by promising goods or resources in exchange for support — what political scientists call clientelistic parties — to win elections. Those parties may have been more likely to rely on ballot fraud to win, the researchers hypothesized, which became harder once electronic voting was introduced.

By contrast, so-called programmatic parties, which tend to mobilize support by promising to enact certain policies — for example the leftist agenda of the Workers Party, the party of the current president, Luiz Inácio Lula da Silva — did slightly better under the new system. (Programmatic parties can have right-wing agendas, too. And no party is exclusively one model or another, but they tend to skew in one direction.)

Nieto-Matiz and Skigin wondered whether there might be a link between violence and the type of party that prevailed in elections, so they set out to test that more rigorously — resulting in the current paper.

They examined a set of local elections so tight that they were essentially coin flips, making the results as close to random as it would be possible to be in real-world politics. That way, they could be reasonably confident that differences were because of the type of party that won, rather than the underlying conditions in the district.

The results were striking: When programmatic parties won, local homicide rates immediately fell. But when clientelistic parties won, violence in their districts actually got worse. And, once again, the researchers said, the effect showed up far too quickly for it to be the result of new laws or policies.

One study isn’t enough to conclusively say why they found a correlation between programmatic parties and reduced violence, the researchers were careful to note when we spoke. But they had a hypothesis — and that’s where things start getting Scorsese-ish.

They suggest that clientelistic parties are more likely to collaborate with local armed groups, which in Brazil include criminal gangs and paramilitary groups backed by landowners and oligarchs.

Research has shown that clientelistic parties tend to have relatively loose internal controls on membership and candidates, which can make them useful vehicles for criminals looking to get into politics — something that other studies have found in India and Colombia. Additionally, gangs and paramilitaries can help get rid of political opposition, assist with election fraud, or deliver the votes of people from groups or areas under their control.

By contrast, because programmatic parties need to maintain ideological discipline, they tend to have stronger institutional controls over who can be a party candidate or official. And they might also face more of a backlash if voters perceive them as corrupt or violent, because their appeal to voters is based on how well they enact their ideological agendas in office. That’s harder to do while mired in investigations or prosecutions for wrongdoing, which means they have less incentive to collaborate with violent groups.

So the theory goes that, while individual politicians’ decisions might vary quite a bit, clientelistic parties had more of an incentive to enter into mutually beneficial relationships with gangs, paramilitaries or other violent actors. And that gave those armed groups more impunity and local power, which in turn increased violent crime.

Which brings us back to the resilience of Brazilian democracy.

Research has shown that over time, programmatic parties tend to crowd out clientelistic parties, because support for the latter tends to collapse as soon as they’re out of power and unable to distribute resources to supporters. Skigin and Nieto-Matiz’s work adds to that by showing how the process might also reduce the power of violent groups.

“We should expect that those that those criminal actors or generally coercive actors, they should be either weakened, or, if they are able to survive, they are not going to be able to resort to as much violence,” Skigin said.

Viewed through that lens, the broader story of Brazil’s democracy starts to look less like an episode of democratic crisis, and more like turbulence on a long, slow and still incomplete trajectory of democratization.

And it suggests that the recent election, which saw the victory of a candidate for the Workers Party — the largest programmatic party in the country — may have implications for ordinary citizens’ lives that go far beyond his party’s policies or ideology.

Thank you for being a subscriber

Read past editions of the newsletter here.

If you’re enjoying what you’re reading, please consider recommending it to others. They can sign up here. Browse all of our subscriber-only newsletters here.

I’d love your feedback on this newsletter. Please email thoughts and suggestions to interpreter@nytimes.com. You can also follow me on Twitter.