This is the first of two parts. Part 2 covers the acute case — the Verfassungsschutz classification of the AfD — the strongest counterarguments, and what the pattern costs.
Introduction
On 26 February 2026, an administrative court in Cologne temporarily suspended the designation by Germany’s domestic intelligence agency, the Bundesamt für Verfassungsschutz (BfV), of the Alternative für Deutschland (AfD) as a “confirmed right-wing extremist” organisation. The ruling was technically limited to the court having found insufficient evidentiary certainty on one specific claim: that the party systematically sought to degrade the legal status of Germans with immigrant backgrounds to second-class citizens. What it did not do was overturn the standing lower-level classification under which the party had been monitored for five years, nor did it address the broader question of whether an executive intelligence agency (BfV) should be making determinations of this kind at all. And it did not lift the material consequences that had flowed from the higher classification for the preceding nine months: restricted banking access, employment pressure on public-sector members, and denied parliamentary security credentials.
What the ruling did was confirm, in the precise technical language of administrative law, that the system works: the executive classifies; the judiciary reviews. Nine months after the higher classification was published, an administrative court found that one specific factual claim was insufficiently evidenced and suspended the designation.
This is judicial review functioning, but it is also judicial review arriving after the damage, in an administrative rather than constitutional court, on the question of whether the BfV’s facts were right, not whether an executive security agency should be making determinations of this kind, with material electoral consequences for the second-largest party in the Bundestag, at all. That gap between the question the court was equipped to ask and the question the situation actually raises is what this essay addresses.
The question is not whether the AfD’s specific combination of positions places it outside the German constitutional order; this essay does not argue the classification itself. What it argues is that the mechanism deserves scrutiny because the mechanism does not confine itself to this case: it is the operational form of an institutional architecture that extends from the German party ban provision through the European Union’s (EU) democratic conditionality to a one-billion-euro-per-year domestic civil society funding apparatus. Every component of that architecture evaluates itself by criteria it produces itself. That is the definition of a closed normative loop, and a closed normative loop is incapable of self-correction.
The intellectual origins of this architecture are not obscure. In 1937, Karl Loewenstein, a German-Jewish jurist who had watched the Nazi party exploit every legal mechanism of the Weimar Republic on its path to power, published two articles in the American Political Science Review arguing that democracies must develop the capacity to defend themselves against movements that use democratic procedures to abolish democratic institutions. He called the framework streitbare Demokratie: militant, or combative, democracy. The framework he designed in Weimar’s ruins became, through the Basic Law (Grundgesetz), the intellectual architecture of post-war Germany; and Germany became, through European integration, the unacknowledged template of the EU’s democratic values system. What began as a precisely targeted, anti-totalitarian instrument, aimed in Loewenstein’s own analysis at fascism and communist totalitarianism alike, drifted across decades of application until it operated almost exclusively against the fascist and far-right pole, carried forward by institutional infrastructures whose accumulated weight made self-examination progressively harder to perform.
The essay that follows traces that drift across three historical phases: an external template established in Portugal’s 1974 transition, extended across Eastern Europe’s 1989 enlargement, and ultimately turned inward against the voters of the state that built it. It argues that the resulting apparatus constitutes a closed normative loop, institutionally incapable of asking whether its definition of democracy is universal or specific, whether the civil society it funds is genuinely independent, or whether the parties it classifies as constitutional threats are genuinely anti-constitutional rather than politically inconvenient. The loop’s most acute current expression, the Verfassungsschutz classification of the Alternative für Deutschland, is both the essay’s central case and its sharpest illustration of what a mechanism looks like when it can receive confirming evidence but not disconfirming. The case for examining it honestly is a case for democratic self-defence remaining credible, not for abandoning it.
II. Loewenstein’s Gift and Its Limits
In the spring of 1937, as fascist movements consolidated power across Europe and liberal democracies struggled to articulate a response that did not look like what they were opposing, Karl Loewenstein went directly to the constitutional core of the problem. The dominant liberal instinct — that democratic procedures protect everyone, including those who seek to use them against democracy — was, he argued, not a principled consistency but a structural vulnerability. A democracy that extended equal procedural rights to movements whose explicit aim was the abolition of those rights was not displaying constitutional virtue. It was, in the precise technical sense, suicidal.
His proposed framework, streitbare Demokratie or combative democracy, had a specific and demanding target, not a licence for the suppression of political dissent. The load-bearing distinction that ran through both articles was between parties that used democratic procedures as tactical cover toward anti-democratic ends, and parties that pursued unpopular or illiberal policies through democratic means. The former constituted a legitimate target of democratic self-defence; the latter, however objectionable their positions, did not. A party that campaigned for the abolition of parliamentary government and then won a parliamentary majority was categorically different from a party that campaigned to restrict immigration, raise tariffs, or nationalise industry and won the same majority. Militant democracy was designed for the first; it had nothing to say about the second. Everything in the framework’s subsequent history depends on whether this distinction can be maintained. The argument of this essay is that it cannot. The failure is not inevitable, but naming it requires first understanding what the distinction actually was.
What Loewenstein was arguing against is equally important. His framework was anti-totalitarian in its structural logic, not anti-fascist in its scope. The theoretical problem, a movement exploiting the open machinery of democratic competition to disable that machinery once in power, applied with equal force to communist parties operating under Comintern direction as to fascist ones. In the same 1937 articles, Loewenstein analysed both; fascism was the historical urgency, totalitarianism the conceptual target. A theory of democratic self-defence that is structurally limited to one form of anti-democratic movement is not a theory of democratic self-defence but a partisan instrument that happens to invoke constitutional language, distinguishable from the genuine article in that one applies consistently wherever the structural threat appears, the other selectively, wherever its beneficiaries find it convenient. The drift of the militant democracy apparatus from the former toward the latter is the argument this essay traces, and that drift can only be identified as drift if the framework’s stated universality is taken seriously at the outset.
The men who drafted the West German Basic Law in 1949 took it seriously. The constitutional settlement they produced was not a set of abstract principles but a targeted response to specific mechanisms through which democratic self-destruction had occurred, designed by people who had watched those mechanisms operate at close range. Article 21(2) encoded Loewenstein’s framework — parties that seek to undermine or abolish the free democratic basic order shall be prohibited — but the wording gave effect only to his central proposition; the institutional architecture gave effect to something equally important that his articles had not specified: who should make the determination.
The Basic Law’s answer was the Federal Constitutional Court (Bundesverfassungsgericht, BVerfG) — neither the Interior Ministry, nor an intelligence agency, nor a parliamentary majority, but the highest court in the constitutional order, operating under the most demanding evidentiary standard, through a full adversarial proceeding with public scrutiny and opportunity for counter-argument at every stage. The founders understood, because the Weimar Republic had demonstrated, that the power to classify parties as threats to the constitutional order is itself a power subject to abuse. A government that could make that determination without an independent judicial gate would possess precisely the instrument for political entrenchment that the Basic Law was designed to deny, which is why the judicial architecture was not procedural nicety but the design’s essential safety mechanism.
The early case law confirmed both the symmetry and the seriousness of the design. In 1952, the Constitutional Court banned the Socialist Reich Party (Sozialistische Reichspartei, SRP), an openly neo-Nazi successor movement that had already won parliamentary seats in Lower Saxony, establishing the foundational jurisprudence: the constitutional order is value-based, not value-neutral; parties that reject human dignity and seek to replace democratic institutions with authoritarian ones are outside the constitutional order regardless of their electoral performance. Four years later, in 1956, the same court banned the Communist Party of Germany (Kommunistische Partei Deutschlands, KPD) under the same provision and the same stated reasoning: a finding that the KPD’s programmatic hostility to democratic institutions, sustained through ideology and organisational practice, placed it outside the constitutional order. Both totalitarian poles subjected to the same judicial instrument: the theoretical intention made operational.
What happened after 1956 was not a reversal but a drift: slow, unremarked in official discourse, and to this day never acknowledged as such. The KPD ban was not revisited when the party re-founded itself in 1968 under a different name with a recognisably similar programme. The Party of Democratic Socialism (Partei des Demokratischen Sozialismus, PDS) — the direct organisational successor of the Socialist Unity Party (Sozialistische Einheitspartei Deutschlands, SED) that had governed East Germany for forty years — participated in federal elections, entered the Bundestag, and formed parliamentary groups without triggering even the preliminary scrutiny that preceded the 1956 ban. Die Linke (The Left), which absorbed the PDS, has sat in state coalition governments. The Desiderius-Erasmus-Stiftung (Desiderius Erasmus Foundation, DES) — affiliated with a party that in 2021 received 12.6 percent of the federal vote and by 2025 was polling in the mid-twenties — is denied the state funding that every other significant party’s foundation receives as a matter of course. The application of militant democracy’s instruments has become, in practice, weighted almost entirely toward the fascist and far-right pole.
The explanation for this drift is genuine but not a justification. In West German political consciousness, the Nazi legacy was immediate, domestic, and morally overwhelming in ways that the GDR’s communist legacy was not. The Holocaust occurred on German soil, perpetrated by German state institutions, within the living memory of the Basic Law’s drafters. The GDR (German Democratic Republic, Deutsche Demokratische Republik), by contrast, was geographically separated — another state, behind a wall, under Soviet occupation — and was frameable, with some plausibility, as an external imposition on a divided nation rather than as a German pathology. The Social Democratic and trade union traditions that dominated West German reconstruction were not natural opponents of communism in the way they were of fascism; the Frankfurt School, the party foundations of the SPD (Sozialdemokratische Partei Deutschlands, Social Democratic Party of Germany), and the Ostpolitik intellectual tradition all emerged from a political culture whose antennae were calibrated primarily to the fascist frequency. Explaining why an asymmetry developed is not the same as establishing that it is analytically defensible, particularly in a framework that explicitly claimed to be applied on structural rather than ideological grounds. An essay that presents itself as a fair-minded application of Loewenstein’s universalist principles is obliged to name the asymmetry before it can be trusted to analyse it.
The reason the naming cannot be deferred is the East German mirror case, which is the most powerful demonstration available that the self-legitimising loop is a structural property of the mechanism rather than a property of any particular ideology operating it. The demonstration’s analytical value lies not in any claim of moral equivalence between the two states — the Federal Republic is a constitutional democracy; the GDR was not — but in the structural identity: the same institutional logic, applied from the opposite ideological pole, produces the same pattern of self-confirmation.
The German Democratic Republic constructed its entire political identity around anti-fascism, not merely rhetorically but as a systematic institutional programme. The SED’s official historiography presented the communist state as the legitimate heir of German resistance to National Socialism, opposed to a West Germany that, in the SED’s framing, had provided shelter and career continuity to Nazi industrialists, Nazi judges, and Nazi officers. There was enough truth in this to give the framing purchase; the Federal Republic’s early personnel continuities were a genuine scandal. But the framing did far more than identify a genuine deficiency in West German denazification. It provided the SED with a closed normative loop that operated exactly as the one this essay traces in the Federal Republic, running from the opposite ideological pole.
Within the logic of the anti-fascist state, any citizen who challenged SED governance was not a political opponent holding a different view but someone objectively aligned with fascism: proto-fascist, enemy-adjacent, a threat to the democratic anti-fascist republic that the GDR claimed to be. The Ministerium für Staatssicherheit (Ministry for State Security, MfS, colloquially the Stasi) classified, monitored, and where necessary imprisoned people whose actual political positions ranged across the full spectrum from liberal democrat to social democrat to reform communist. A citizen who objected to housing allocation, attended an unlicensed church gathering, or applied to emigrate could find each of these ordinary acts recorded in a Stasi dossier as evidence of feindlich-negative Kräfte (enemy-negative forces) whose existence the subject would not learn until the archives were opened after 1990. The language of democratic protection against fascism was the institutional cover under which this occurred. Challenges to the mechanism were absorbed as confirmation of the threat: if you questioned whether the SED’s framework was genuinely anti-fascist, you were demonstrating fascist sympathies. The loop was closed from inside.
The structural parallel to what this essay traces in the Federal Republic is exact. A state defines democracy: in the GDR, through the anti-fascist framework; in the Federal Republic, through the freiheitlich-demokratische Grundordnung (FDGO, the free democratic basic order) and the civil society ecosystem that embodies it. The state empowers actors who operationalise that definition: the SED’s approved mass organisations; the Federal Republic’s party foundations and Demokratie leben! grantees. The state classifies those who fall outside the definition as threats: the Stasi’s feindlich-negative Kräfte; the Verfassungsschutz’s gesichert rechtsextremistisch. Challenges to the mechanism are absorbed as confirmation of the threat: in the GDR, any challenge demonstrated fascist sympathy; in the Federal Republic, the AfD’s legal challenges appear in BfV classification materials as evidence of the party’s resistance to democratic norms, and thus as further evidence for classification.
The ideological poles are reversed but the structure is the same. This is not a claim of moral equivalence: the Federal Republic’s independent courts, constitutional guarantees, and free elections mean that a citizen who challenges the democratic promotion apparatus here does not face the consequences that challenge invited in the GDR. But equivalence is not the argument. If the self-legitimising loop was self-legitimising when the SED operated it, and it manifestly was, then the question of whether it is self-legitimising when the Federal Republic operates it cannot be answered by pointing to the difference in ideology but only by pointing to the difference in mechanism: specifically, the presence or absence of the judicial gate that the Basic Law’s founders correctly identified as what distinguished a constitutional defensive instrument from a political weapon. The GDR had no independent judiciary capable of constraining the SED’s use of the anti-fascist framework. The Federal Republic does. More precisely, it does in the domain where the Basic Law’s designers placed the safety mechanism. Whether that safety mechanism extends to the domains where the apparatus has since migrated is the argument’s central question.
The most recent episode in the party ban mechanism’s own history provides a partial answer, revealing the judicial mechanism functioning precisely as the founders intended.
In 2017, the Federal Constitutional Court declined to ban the National Democratic Party of Germany (Nationaldemokratische Partei Deutschlands, NPD), a party whose neo-Nazi character had been acknowledged by multiple courts, whose rhetoric included explicit ethnic nationalism, and whose organisational connections to post-war fascist networks were documented in detail. The reason was not insufficient evidence: the Court found that the NPD acted “systematically and with sufficient intensity” toward aims directed against the free democratic basic order. What it found insufficient was the realistic prospect that the NPD could translate those aims into actual constitutional threat. At 1.3 percent of the federal vote, the party had no realistic path to power, and the Court introduced what has since become known as the potentiality criterion: anti-constitutional intent, however thoroughly documented, does not suffice for a ban without realistic capacity to constitute an actual threat to the constitutional order.
This was the judicial mechanism performing exactly the self-corrective function the founders designed for it. The Court recognised that the party ban instrument applied without a proportionality check — used against parties whose ideology is objectionable but whose capacity to threaten democracy is minimal — begins to look less like constitutional self-defence and more like elite suppression of unpopular political views, and that this appearance is itself a threat to democratic legitimacy. The potentiality criterion is an internally generated constraint on overreach, developed through seventy years of jurisprudential evolution.
The Verfassungsschutz (constitutional protection service) classification mechanism has developed no equivalent. Its procedure includes no potentiality criterion, no requirement that the classified party have realistic capacity to constitute a constitutional threat proportionate to the classification’s consequences, and no public adversarial proceeding in which evidence is tested against counter-argument under judicial scrutiny. What it has instead is an executive agency, reporting to the Interior Minister, making determinations that carry the practical consequences of quasi-judicial rulings without the architecture that makes such rulings defensible.
The migration from the judicial to the executive — from the mechanism that Article 21(2) designed to the one that grew up alongside it, in the unmonitored space the Basic Law did not explicitly address — is not a footnote to the militant democracy story but its decisive development. What follows traces where that migration led, what it produced, and what it cannot see about itself.
III. The Operational Sequence: Three Phases
The migration from judicial home to executive and administrative space happened not through a single institutional decision but across three decades and three distinct phases of application: each extending the apparatus into new territory, each generating justifications that made the extension appear as natural continuation of the last, and each building infrastructure whose accumulated weight made self-examination progressively harder. The phases are historically distinct; the logic running through them is continuous.
Phase One: Portugal and the Template (1974–1986)
On 25 April 1974, a military coup in Lisbon ended forty-eight years of authoritarian rule in Portugal. The officers who organised the Carnation Revolution had no unified political vision beyond the removal of the Estado Novo regime. What emerged from the coup was not a democratic transition but a political vacuum: contested simultaneously by competing military factions, a Communist Party that had survived the Salazarist decades underground and had organisational capacity that no other Portuguese political force could match, and a collection of democratic parties that had been effectively suppressed for two generations. The Portuguese Communist Party, under Álvaro Cunhal, was aligned with Moscow, opposed to parliamentary democracy, and positioned to convert the post-coup chaos into a revolutionary outcome. Whether it would succeed was, in the eighteen months following April 1974, a genuinely open question.
Into that opening stepped the Friedrich-Ebert-Stiftung (Friedrich Ebert Foundation, FES) — the SPD’s party foundation — alongside the Socialist International, a network of European Social Democratic parties operating under a clear political brief. Their vehicle was Mário Soares’s Portuguese Socialist Party, founded in 1973 in West German exile. The intervention was neither secret nor modest: financial support, organisational training, political networking across Western European capitals, and the explicit backing of the Socialist International gave the PS a structural advantage it could not have assembled from domestic resources. The West German SPD, under Willy Brandt, treated the Portuguese transition as a strategic priority: a Portugal that fell to communist governance would not join the European Community; one that achieved Social Democratic consolidation would, and would also provide a model for Spain, where Franco’s succession was already in prospect.
Soares’s PS won the 1975 constituent assembly elections with 38 percent of the vote. The Communist Party’s peak was contained. A democratic constitution was adopted, and Portugal joined the European Community in 1986, its consolidation institutionalised by the accession framework. By any reasonable measure, the intervention had produced a good outcome. That is precisely the analytical problem.
The template that Portugal established was not that external support for democratic transitions is useful. It was something more specific: external political intervention in a neighbouring state’s democratic formation is legitimate if directed toward liberal democracy, and what the Friedrich Ebert Foundation pointed toward was not democracy in the abstract but specifically Social Democratic democracy, organised against both the communist left and the authoritarian right. The values encoded in the intervention were not neutral democratic procedure but the values of a specific political tradition, justified by the outcome that tradition produced. Portugal’s democratic consolidation meant EC membership, which meant economic integration, which meant the outcome retrospectively validated everything the intervention had done.
This is the template’s self-legitimising quality in its first instance: a successful democratic transition validates the intervention that shaped it, making the outcome evidence enough that non-neutrality need not be examined. The circular reasoning is invisible from inside: the intervention produces democracy; democracy proves the intervention was democratic. From outside, as Eastern European governments would later try to articulate, it looks rather different: powerful external actors shaped which political forces prevailed in Portugal’s transition and which were marginalised, using their own values as the standard, and declared the result democratic because it matched those values. Whether it was genuinely self-determined, democratic in a procedural rather than substantive sense, was a question the template had no mechanism for asking, and would not ask. The toolkit — party foundations, Socialist International networks, civil society support, accession conditionality — was ready for the next transition.
Phase Two: Eastward Extension (1989–2004)
The fall of the Berlin Wall presented the toolkit’s operators with what looked like the largest democratic promotion opportunity in European history. Across Eastern Europe, communist regimes were dissolving, some rapidly, some violently, some through managed negotiation, and new democracies were forming in conditions of institutional vacuum, economic shock, and social fragmentation. The Western European institutions that had developed during the Cold War — the party foundations, the EU accession framework, the conditionality architecture — were ready for exactly this moment. The Copenhagen Criteria of 1993, which formalised the conditions of EU accession as democratic governance, rule of law, human rights, and market economy, were the template institutionalised at European scale.
What their success in Portugal had trained the toolkit’s operators not to examine was whether the conditions in which the toolkit had worked — a Southern European Catholic country with a specific colonial past, Salazarist institutions, and a Social Democratic tradition seeking integration into a Western European economic framework — bore sufficient resemblance to the conditions in which it would now be applied, which they did not. The Eastern European states had experienced fifty years of Soviet-directed communism that had systematically destroyed independent civil society, market institutions, and organic political formation. They had pre-communist traditions, national, religious, and aristocratic, that the communist period had suppressed but not dissolved, and which re-emerged after 1989 in forms that Western European political culture struggled to categorise. They had relationships to sovereignty, minority rights, and national identity shaped by experiences — partition, occupation, population transfers, ethnic cleansing — that the Western European founding generation had largely shared but had resolved, through the EU project itself, by placing sovereignty pooling at the centre of the answer.
The civil society built through EU and foundation support in Eastern Europe was structurally different from that which had emerged organically in Western European transitions: built from outside in, through grant cycles and programmatic frameworks designed for Western European political assumptions, staffed by professionals trained in Western European NGO culture, oriented toward Western European standards of democratic achievement. Ivan Krastev and Stephen Holmes, in their 2019 study of Eastern European liberalism’s crisis, identified the dynamic precisely: Eastern European elites were expected to implement policies they had not designed, through institutions they had not built, to achieve outcomes defined by standards they had no part in setting. The result was mimicry, not ingenuity: a performance of democratic consolidation evaluated against Western criteria, by Western-funded organisations, for Western audiences.
Hungary and Poland were the template’s greatest success stories — both negotiating accession with speed and competence, both celebrated as proof that democratic conditionality produced durable consolidation. By 2010, Hungary’s Fidesz — led by Viktor Orbán, himself a product of the Soros-funded Colleges for Advanced Studies that had been exactly the kind of civil society investment the template prescribed — had begun systematically dismantling the independence of courts, media, and electoral institutions while retaining the formal apparatus of parliamentary democracy. Poland’s Law and Justice party followed an analogous trajectory from 2015.
The EU’s institutional response is the template’s self-confirming quality in its starkest form: Hungary and Poland were not understood as disconfirming evidence — as evidence that the template’s assumptions about how democratic consolidation works were wrong, or at least incomplete — but as pathological deviations, democratic backsliding, the work of illiberal leaders exploiting their populations’ democratic deficits. The solution was more conditionality, more rule-of-law proceedings, more civil society support. The template was the answer to the template’s failure, which is precisely the property of a closed normative loop: it can receive only evidence requiring further extension, never evidence requiring revision.
The question the framework cannot ask is whether Orbán’s government, whatever its genuine democratic deficiencies, was expressing something about Hungarian political culture and historical formation that the template had failed to engage rather than merely overcome; whether the legal independence of Hungarian courts, built rapidly from external specifications in the 1990s, was the same thing as judicial independence rooted in historical legitimacy; whether Polish national conservatism, shaped by centuries of partition and Soviet occupation, had legitimate reasons to resist certain forms of EU conditionality that were not simply reasons for further conditionality. The template knows how to classify these positions as democratic backsliding. It has no vocabulary for asking whether they are also something else.
Phase Three: Domestic Institutionalisation (2015–present)
The third phase is structurally identical to the first two. The same logic — state-backed, ideologically specific democratic promotion, justified by appeal to democratic protection — operates in the same institutional forms. What changes is the target population. In Portugal it was Portuguese citizens; in Eastern Europe it was Polish and Hungarian citizens; in Phase Three it is German citizens in Saxony and Thuringia. The moral distance that made the non-neutrality easy to overlook in the earlier phases has contracted to nothing. The apparatus is now domestic, operating against the voters of the state that funds it.
The German democratic promotion apparatus totals approximately one billion euros per year across three institutional levels that are parallel and mutually reinforcing rather than independent. The party foundation system — six foundations, each affiliated with a parliamentary party, together receiving approximately €660 million per year, roughly ninety-five percent from state sources — is the structural bridge between the external phases and the domestic one. These are the institutions that trained the political cadres, built the civil society organisations, and delivered the international legitimacy of the Portugal and Eastern European phases; having completed that work externally, they turned the same toolkit toward domestic German civil society: scholarship programmes, academic networks, political training, publications, research institutes, and direct grants to organisations operating in their ideological orbit. The Bundeszentrale für politische Bildung (bpb), the Federal Agency for Civic Education with approximately 420 staff and a budget of €103 million per year, produces the textbooks, digital resources, and policy analyses through which a generation of German students encounters political education, operating without editorial independence from the Interior Ministry portfolio and incorporating Verfassungsschutz classifications into its educational output as authoritative content: the security agency classifies the threat; the civic education agency teaches it as fact. The Demokratie leben! (”Democracy Lives!”) programme, at €191 million per year funding 600 to 800 civil society organisations at any given time, is the direct civil society funding layer, with an ideological register specific enough that anti-racism, diversity, gender, migration, and countering right-wing extremism are its dominant frames. The Correctiv investigative journalism collective, which in January 2024 published a report on a meeting of AfD-adjacent figures discussing population remigration that produced the largest street demonstrations in German history, received funding through this ecosystem: an instance of state-funded political activity intersecting with the electoral cycle in a specific directional way.[^1]
Together these three levels constitute a mechanism by which the state funds, trains, informs, and operationally supports a civil society that embodies a specific definition of democratic culture, and then uses that civil society’s assessments to evaluate whether Germany is adequately democratic. Each component has a defensible purpose in isolation. As a system they form a loop without external corrective.
The Desiderius Erasmus Foundation makes that loop visible in its purest form. The AfD is the second-largest party in the Bundestag, and the criteria by which party foundations receive state funding, consistent parliamentary representation across two electoral cycles, formally entitle the DES to approximately €25 million per year. It has received effectively nothing since its founding in 2017, operating on approximately €200,000 in private donations.[^2] The formal criteria exist but are overridden by a substantive political judgment — that the DES does not stand on the basis of the free democratic basic order — derived from the same Verfassungsschutz classification examined in Part 2. The mechanism that determines which parties may have state-funded civil society infrastructure is the same executive agency mechanism that classifies parties as constitutional threats. The Rosa-Luxemburg-Stiftung (Rosa Luxemburg Foundation) — affiliated with Die Linke, the direct institutional successor to the party that governed the GDR — receives €74 million per year on the same formal basis; it is inside the loop while the AfD is outside it, and the boundary between the two is drawn by an executive agency without judicial oversight.
In March 2026, CDU (Christlich Demokratische Union, Christian Democratic Union) Education Minister Karin Prien announced that approximately 200 Demokratie leben! projects would lose funding by year-end.[^3] Her stated rationale was empirical: the funded programmes were not demonstrably achieving their stated objectives. Her proposed alternative was modest: shift funding toward embedded institutions, schools, libraries, fire brigades, sports clubs, with stricter pre-review criteria. The response from the ecosystem she proposed to restructure was not empirical. SPD spokeswoman Jasmina Hostert warned that terminating the projects risked “misallocations to the detriment of democracy,” treating the funded programme as constitutive of democracy rather than instrumental to it. Left Party representatives framed the restructuring as fulfilling the AfD’s “boldest dreams.” Green MPs accused Prien of “general suspicion against civil society.”
The rhetorical move is uniform and revealing: the defenders did not argue that the programme was effective and worth preserving on evidential grounds but that questioning its ideological balance was itself anti-democratic, that the programme and democracy had become indistinguishable, so that reform of the one was attack on the other. The Swiss Neue Zürcher Zeitung, observing from outside the loop’s orbit, identified the circular logic explicitly: critics of Prien’s Verfassungsschutz integration were themselves embedded in the Demokratie leben! ecosystem, which meant their critique of state security involvement in democracy promotion was structurally indistinguishable from their defence of their own funding stream.[^4] The circularity was visible from Zurich in a way it was not from Berlin.
This is Phase Three’s completion. The loop that began in Portugal — external intervention in a neighbouring state’s democratic formation, justified by the direction of that formation toward liberal democracy — has turned inward and found its own citizens. The state defines democratic culture, funds the civil society that embodies it, uses that civil society’s assessments to evaluate democratic health, and responds to challenges to this circuit by designating the challenge itself as a threat to democracy. The most acute current expression of that designation is the subject of what follows.
Notes
[^1]: Correctiv received approximately €411,147 from the Demokratie leben! programme from 2023 onward — comprising €280,991 from the Federal Family Ministry for the project ‘Brandherd Desinformation’ — with additional NRW state funding bringing the combined total to approximately €460,000. The Federal Research Ministry separately funded Correctiv with €720,418 in 2026. See ‘Correctiv erhält knapp 460.000 Euro aus dem Programm Demokratie leben! und dem Land NRW’, Die Welt / MSN Deutschland, 17 May 2026.
[^2]: The DES’s operational dependence on private donations of approximately €200,000 annually, against a formal entitlement of approximately €25 million under party foundation criteria, is documented in Ralf Fücks, ‘Will Germany Be Forced to Fund the AfD’s Desiderius Erasmus Political Foundation?’, Foreign Policy, 15 June 2022; and in the Verfassungsblog analysis ‘(Kein) Geld für die Desiderius-Erasmus-Stiftung: Zur Frage, ob die AfD-nahe Stiftung 2026 auf Fördermittel in Millionenhöhe hoffen kann’, Verfassungsblog, 2026. The criteria for party foundation funding are set out in the party foundation guidelines (Richtlinien für die Förderung der politischen Bildungsarbeit der Stiftungen).
[^3]: The Prien announcement is documented in Jonas Hermann, ‘Steuergeld nur für politisch genehme NGO? Besser, man streicht allen die Förderung’, Neue Zürcher Zeitung, 5 March 2026; ‘Demokratie leben: NGOs — Prüfung’, Nius.de, March 2026; and subsequent national reporting in Der Spiegel and Süddeutsche Zeitung. The parliamentary responses cited — SPD, Greens, Left Party — are from debate coverage in the same period.
[^4]: Nathan Giwerzew, ‘«Demokratie leben!» und der Verfassungsschutz: Der Irrweg der Karin Prien’, Neue Zürcher Zeitung, ‘Der andere Blick’, 19 June 2026. Giwerzew identifies the structural problem: critics of the proposed Verfassungsschutz integration into Demokratie leben! applicant screening are themselves embedded in the programme’s ecosystem, making their objections structurally indistinguishable from a defence of their own funding.
Part 2 continues with the acute case — the Verfassungsschutz classification of the AfD — the strongest counterarguments the essay must answer, and the costs of the pattern traced here
In-depth article; thanks for the effort!
The same mechanism has been at work in the Netherlands for some time.
Pay particular attention to the term ‘Anti-institutioneel extremisme’ (anti-institutional extremism)
Update, September 2026 — three things since this essay went up that are well worth mentioning:
Firstly: The choice the essay describes as *structural* has now been stated outright: interior Minister Dobrindt's formula for handling the AfD, repeated since May 2025 is "Nicht wegverbieten, man muss sie wegregieren" — don't ban them away, govern them away. Chancellor Merz went further, warning that an outright formal ban could look like "the political elimination of a competitor."
Secondly: the AfD's youth wing, the "Junge Alternative" (Young Alternative), was itself confirmed extremist by the Cologne court. This is a stronger categorisation than the federal party has sustained and AfD's response wasn't to contest it but to dissolve the organisation and relaunched it as "Generation Deutschland." (German Generation)
Arguably this is another example of a closed loop that doesn't correct itself and it needs following up on as it develops.
Thirdly, the scale of this: national polling now has the AfD at 27.9%, ahead of every other party. The essay's question about what streitbare Demokratie means at this size is becoming very relevant