In December 2024, Romania’s Constitutional Court did something almost without precedent in modern European history: it annulled the first round of a presidential election 48 hours before the second round was scheduled after intelligence reports revealed a Russian-backed disinformation campaign had distorted the result. The candidate who had won that round, Călin Georgescu, was subsequently barred from the rerun entirely and now faces criminal charges, including incitement to overthrow the constitutional order.
Romania’s institutions had, from one perspective, defended democracy. From another, they had just demonstrated exactly what democratic erosion looks like. Both views are partially correct, which is precisely what makes Romania’s crisis so important and yet so difficult.
The TikTok Threat
Georgescu was polling at roughly 5% just weeks before the November 2024 first round, yet he won it with 23%. The explanation for that surge lies not in a sudden shift in Romanian political opinion but in a coordinated infrastructure of manipulation. Romanian security services documented over 85,000 cyberattacks on electoral IT systems, bot networks amplifying Georgescu’s content across Telegram and TikTok, and AI-generated material that accumulated approximately 150 million views in two months. Stolen server credentials were found on Russian forums, and intelligence assessments linked the campaign directly to Russian hybrid operations. Georgescu himself reported spending zero euros on his campaign, but Romanian businessmen estimated the true cost at €50 million.
This is what scholars of democratic backsliding increasingly identify as the frontier of electoral manipulation: not the older practices of ballot-stuffing or voter intimidation, but the systematic distortion of the information environment on which democratic decision depends. As Steven Levitsky and Lucan Way argue in their work on competitive authoritarianism, elections can be formally free while being substantively unfair when the playing field is so tilted by manipulation that genuine competition no longer exists.
Romania’s first round was not stolen in the traditional sense. Rather, it was engineered.
Cure and Complications
The Constitutional Court’s annulment was, under this analysis, a defensible act of what scholars call “militant democracy”—the principle, developed by political theorist Karl Loewenstein in the 1930s, that democracies must be willing to use their own legal mechanisms to defend themselves against actors who seek to exploit democratic openness to destroy it from within.
The Court did not act arbitrarily. It cited overwhelming evidence of interference and invoked principles of electoral integrity that are foundational to democratic legitimacy. But the annulment created another issue when it was announced. Georgescu’s supporters—and, quickly, far-right movements across Europe and the United States—reframed the entire episode as proof that liberal elites rig elections when they dislike the results. JD Vance told the Munich Security Conference in February 2025 that Romania had cancelled an election “based on the flimsy suspicions of an intelligence agency.” His framing spread across the same social media networks that had manufactured Georgescu’s rise in the first place.
This is the core paradox of militant democracy: the tools used to defend democratic institutions can simultaneously generate the populist narratives that erode trust in those same institutions.
Nancy Bermeo’s framework of “executive aggrandizement” is instructive here, but in an unexpected direction. The threat in Romania came not primarily from an executive seeking to expand power, but from an external actor engineering the conditions.
Misconstrued Framing
The dominant framing in Western commentary has been to treat Romania as either a success story (institutions held) or a cautionary tale about judicial overreach. Both miss the more important point: the annulment did not resolve the underlying conditions that made Georgescu viable, it only removed him from the ballot.
In the rerun election held in May 2025, Georgescu’s far-right ally George Simion won 41% of the first round vote, channeling the same base of voters, the same anti-EU and anti-establishment grievances, the same social media infrastructure. The pro-European independent Nicușor Dan ultimately won the second round with 54%, and EU leaders celebrated, but Simion’s 46% was a structural indicator.
Pippa Norris’s analysis of populism applies directly here. Populists can lose elections and still win the culture war. The conditions that produced Georgescu (e.g. economic frustration, institutional distrust, a media ecosystem vulnerable to foreign manipulation) were not addressed by an annulment or a criminal investigation, remaining in place to this day.
What Romania Reveals About the New Threat to Democracy
Romania’s crisis is not primarily a story about one candidate or one court decision. It is a stress test for a broader question that democratic theory has not yet answered: how do democracies defend themselves against threats that are simultaneously external and domestic, technological and political?
The Russian interference in Romania was real. So was the domestic discontent that made that interference effective. Blaming one without accounting for the other produces incomplete diagnoses and ineffective responses.
The institutional response—annulment, criminal charges, a rerun—addressed the external manipulation, but did not address why a candidate promoting pro-Russian views, fascist historical revisionism, and vaccine conspiracy theories was genuinely the preferred choice of nearly a quarter of Romanian voters in the first place.
This demonstrates how democratic erosion does not rest in court decisions, rather in the accumulated failure of institutions to deliver accountability, reduce corruption, and give citizens reasons to trust the system that is meant to work for them.
Why This Matters Beyond Romania
Romania is an early and unusually well-documented instance of a challenge that democracies across Europe and North America are now confronting: the vulnerability of open information environments to coordinated foreign manipulation, and the legitimacy costs of the legal tools used to fight back.
The 2026 congressional midterms in the United States, as well as elections in Germany, France, and Poland, will all take place in information environments where the same infrastructure of bots, algorithmic amplification, and AI-generated content that elevated Georgescu can be deployed again. Romania’s institutions held in 2025, but whether they held for reasons that will convince the 46% who voted against them is the question that only the next election will answer.

0 Comments