Can political culture really protect or doom democracy?

I began the dissertation with a question about courage. People give their lives in the struggle for democracy. Who is willing to risk theirs for free and fair elections, and how much does it matter for whether a democracy is born, or survives, when some of them do?

When I wrote those lines in 2016, fatigue with liberal democracy was rising, yet the most influential theories of how democracies are born and how they die gave little reason to think that faltering public support would matter. Many political scientists doubted that political culture mattered at all, and some evidence suggested that democracy breeds democrats more than democrats breed democracy, even as policymakers went on explaining corruption and failed democracies by culture.

Hume’s point is that even the most military government depends on the opinion of the people it governs, because force is always on their side. Political scientists have spent decades modelling regimes as bargains between elites, armies and economic interests, but popular opinion keeps breaking into those bargains, through at least four channels.

None of these channels needs culture to be fixed or ancient, only that what citizens want, and how hard they act on it, can tip the outcome when a regime is under strain. The same holds inside dictatorships. Scholars of authoritarian rule usually study four survival strategies, and mass support and political culture enter each of them.

By political culture I mean the deep-seated values and practices that shape what citizens demand in politics and how hard they press for it. Culture, in this sense, is information learned from other people, by teaching and imitation. People copy the majority and they copy people of standing. That gives institutions a way in: by making visible how others behave, and by deciding who gets resources and status, institutions can change what people learn. The essay I wrote on this after the 2016 American election, The return of political culture, sets out the argument at more length.

Cross-national studies had found strong correlations between cultural traits such as individualism, egalitarianism and trust, and democratic, clean government. But they compare countries at one point in time, and democracy and clean government themselves raise trust and support for democracy. To get around that, I split the big question into links that can be studied one at a time, in settings that hold more of the rest still. The dissertation’s own overview of its chapters, redrawn below, has three corners: national institutions, political culture and subnational regimes.

Across the designs, the culture people inherited explained less of democracy and political behaviour than the deep-history view suggests. Each panel below has its own scale: a national or inherited difference in grey, and what remains of it when people share institutions, in blue.

Each finding has its own edition:

• Christian and Muslim regions are similar moves the debate about Islam and democracy from countries to 492 African regions.
• Getting to Denmark follows the border through two changes of rule and three kinds of evidence.
• Nordic trust may be younger than it looks, written with Peter Thisted Dinesen, places the Nordic countries in Europe and in history.
• Where you live matters more than where you grew up, with Dinesen as lead author and published in Political Geography in 2022, follows immigrants and their children across Europe.

Only the Africa chapter tests political culture against democracy directly, and it measures democracy as people perceive it, because no one had measured it for every region. Yet regimes vary a great deal inside countries, in how free elections are, how much the press can say and how freely officials repress. Do those differences follow differences in political culture?

Regions that run regimes unlike their country’s are often culturally or linguistically distinct, have a history of independence or autonomy, and lie on the geographical edge of the country. All three hold for Tamil Nadu, though not for Bihar or Sverdlovsk. Liesbet Hooghe and Gary Marks call such places Rokkan regions, after Stein Rokkan’s work on centre and periphery. When populous, they often set off a race to decentralise: their autonomy prompts other regions to demand the same, which gives them still more room to run regimes of their own.

Without systematic data on democracy below the national level, these cases cannot settle the question, because they were picked out for their unusual regimes. What they can show is what political culture looks like inside such regimes. Two of the three I studied most closely were electoral authoritarian enclaves inside democratic India, and the third a democratic enclave inside authoritarian Russia. The Indian enclaves had to win votes, even on a tilted playing field, so legitimacy, branding and propaganda loomed large.

Local political culture can only explain a region’s regime if local actors decide it. Chapters 5 and 6 ask when they do. In Bihar and Sverdlovsk, strong local coalitions took their regions away from the national regime type, against national actors who tried and failed to stop them. Years later, national actors brought both back into line: in Bihar through President’s Rule and the Election Commission in 2005, in Sverdlovsk through Vladimir Putin’s centralisation around 2004.

Chapter 6 adds how such transitions happen. Asking who leads, where their power comes from and whether they break with the national state gives eight modes of transition. The four led by local incumbents, local oppositions or local insurgents look a good deal like transitions in whole countries. The four that run on power from outside the province have few national parallels. Edward Gibson had written that subnational democratisation “is not democratization in short pants”; for transitions led by local actors, I found that it often looks much like it.

Together the two chapters suggest a rule of thumb, in democracies and autocracies alike: transitions away from the national regime are less likely to be led by the centre, and transitions back to it more likely. Where local actors lead, local political culture has room to matter; where the centre leads, it matters less. The editions Regime change in miniature and Transitology in short pants tell both stories.

Most of what I found shows institutions and settings reshaping culture. It does not follow that the culture they produce is irrelevant. A democracy whose citizens have learned to vote, organise and trust each other is harder to dismantle than one whose citizens have not, and a dictatorship that has taught its subjects to see it as legitimate needs less repression.

The practical implication is to study how culture changes, not only how it persists. Migrants in Europe are changed by the people they live among, above all when they do not live in enclaves of their own. In North Schleswig after 1920, it was not only the Danish administration that returned. Danish newspapers and civic organisations that German rule had kept out came back too, and with them the habits of the rest of the country.

These conclusions are narrower than the grand question. They concern political socialisation and immigration, border changes and the incorporation of new citizens, and government below the national level. They are also consistent with each other, and they all cut against the idea that a people’s political culture is a fixed inheritance that decides whether it can be democratic.

The years since have supplied plenty, above all in the United States, where a president refused to accept an election defeat in 2020, his supporters stormed the Capitol on 6 January 2021, and he returned to office in 2025. Much of what decided those weeks was behaviour: officials who certified results, legislators who counted the votes, and people who tried to stop the count. That is political culture in the sense used here, and it can be measured. The essay’s edition has a short afterword with the sources.

My findings speak to several arguments that have only grown louder since. For established democracies they are a warning, because what can be learned can be unlearned: fifty-six years of German rule left North Schleswig voting 10 to 12 points less than its Danish neighbours when it returned. For debates about Islam and democracy, the gap between countries is real, but ordinary Muslims’ political culture does not explain it; states, their histories, their alliances and their treatment of religious minorities do more of the work.

For reformers who want to get to Denmark, they are encouraging: Danish trust looks younger than it seems, and the border evidence suggests that what institutions built in two generations, others can build. For debates about immigration, the evidence is that newcomers’ political habits move towards those of the people around them, slowly in the first generation and much further in the second.

I now work with governments, funders and nonprofits on what transformative AI is likely to mean for low- and middle-income countries. The habit of mind I take from this dissertation is to look for what institutions can build within a generation, and to be sceptical of explanations that treat a society’s capacities as fixed by its history.