The report shows that many citizens define democracy in electoral terms, a very thin definition. Competitive elections and respect for results matter enormously, of course. But judicial independence, legislative oversight, a free press, and other checks on executive power are less central to how many people define democracy.

That helps explain one of the tensions in the report: rising support for democracy can coexist with support for illiberal majoritarianism. People may strongly support elections and majority rule while also tolerating elected leaders who weaken checks and balances, the judiciary, or the press. That really worries me. So the question isn’t simply whether support for democracy is rising, but what kind of democracy people are supporting. 

Then there’s the question of performance. I found Colombia particularly interesting. It saw the region’s largest increase in support for democracy, 12 points, despite sharply deteriorating security, including the highest number of homicides in a decade and a dramatic increase in kidnappings, as well as serious problems in the health system. An increasing number of Colombians, even those with insurance, can’t get access to medicine or appointments.

That complicates the longstanding assumption that support for democracy simply falls when democratic governments fail to deliver.So over the next couple of years, I’d watch three things: whether this increase in support holds, what kind of democracy people are actually supporting, and whether the two foundations emphasized in this report, good governance and inclusive growth, improve.

I understand completely why someone would reach that conclusion. The improvement in daily life is real, and we have to take it seriously. After years of living with extraordinarily violent gangs, Salvadorans can walk through neighborhoods that were once extremely dangerous. You can’t answer that by simply telling people that institutional constraints matter more.

The question is whether weakening democratic institutions is necessary to produce security, and whether that kind of security is sustainable. There’s a difference between a dramatic security improvement and lasting security. Prolonged emergency powers, mass incarceration, and due process violations weaken the rule of law and the institutional capacity that durable security ultimately requires.

There’s also an economic sustainability question in El Salvador. As security becomes a less immediate concern, issues like the employment, poverty, and economic opportunity become more important. The report starts from the proposition that democratic commitment rests not only on good governance, but also on inclusive growth. Extreme poverty has increased in El Salvador, and about 70 percent of workers are in the informal sector, meaning they don’t have a stable income or social protections. Over time, the durability of Bukele’s support is likely to depend in part on whether economic gains reach ordinary Salvadorans.

And we should be very careful about treating El Salvador as a model that can simply be exported. El Salvador’s gang structure and political context are unusual. Countries like Ecuador, Colombia, and Mexico face more complex criminal organizations and transnational illicit economies, which present a very different challenge. The Bukele model won’t work there.

The wrong lesson for leaders to take is that they need to remove institutional constraints in order to get results. The lesson should be that democratic governments have to get much better at delivering security.

They also have to show people what’s working. That means identifying governments that combine effective law enforcement within a rule-of-law framework with targeted prevention, functioning judiciaries, and prison systems that don’t serve as bases of operation for criminal networks.

Independent courts have to prosecute criminals. Security forces have to be both effective and accountable. Democratic leaders have to make the case through performance.

People need to experience democratic institutions as part of the solution, not simply as constraints on government action. Ultimately, the challenge is showing that you don’t have to weaken democracy to deliver security, and that strong institutions are what make security gains last.

I agree with Oliver [Stuenkel] that we should be cautious about calling this a durable ideological realignment. The fact that right-wing self-identification has moved only to 36 percent, despite some pretty dramatic electoral outcomes, is important.

But I’m also not sure anti-incumbency tells us the whole story. There has been a real change in what voters are demanding, particularly around security and order. Right now, candidates on the right have often been more successful at convincing voters that they have answers to those problems.

That doesn’t necessarily mean voters have become much more conservative in a traditional left-right sense. And it connects directly to the conversation we just had about Bukele.

You can have voters who don’t identify strongly with the right but who are increasingly willing to support a strong executive, tougher security policies, and fewer institutional constraints if they believe those things will produce results.

So I think the electoral volatility is partly about performance and anti-incumbency. But underneath that, there may also be a shift in what voters are willing to accept from government in exchange for performance, and our traditional left-right measures don’t fully capture that.

Look beyond presidents and national governments, because mayors and governors are often where citizens most directly experience whether democratic government can deliver security, services, and opportunity, and where confidence in democracy can be rebuilt.